Materials Analysis Applications Compendium • 2017

FTIR • Raman • NMR • NIR • Extrusion • Rheology

 Contents
FTIR Analysis of artists’ pigments by far-infrared microspectroscopy 3

Attenuated total reflection FTIR imaging of “soft chew” formulations 7

Protein secondary structure elucidation using FTIR spectroscopy 10

NIR A case study of using FT-NIR for pharmaceutical hot melt extrusion process monitoring 14

NMR Measuring the equilibrium constant of a keto-enol tautomerism using benchtop NMR 20

Benchtop NMR combined with GC/MS confirms identity of forensic case sample 23

Determination of polymer molecular weight and composition using

picoSpin NMR spectroscopy 28

Qualitative and quantitative analysis of the polymerization of PS-b-PtBA

block copolymer using picoSpin 80 NMR 32

Analysis of acetanilide herbicides and their rotational isomers by picoSpin 80 NMR 36

Hydrogen content determination by picoSpin NMR 40

Evaluation of fluoroorganic compounds with benchtop 19F NMR 44

Raman Classification of polyethylene by Raman spectroscopy 48

FT-Raman: an invaluable addition to the forensic arsenal to combat the opioid epidemic 54

Leveraging the lateral spatial resolution of a confocal Raman microscope to resolve

micron to sub-micron layers in polymer laminates 57

FT-Raman mapping of multicomponent solid dosage forms 61

Characterizing amber and its imitations with dispersive Raman spectroscopy 64

Confocal Raman analysis of a transdermal nicotine patch by a DXR2 Raman Microscope 67

In situ density determination of polyethylene in multilayer polymer films using

Raman microscopy 71

RheoRaman Simultaneous rheology and Raman spectroscopy during the melting

and recrystallization of polypropylene 77

Investigating cocoa butter crystallization using simultaneous rheology and

Raman spectroscopy (RheoRaman) 82

Other Raman and FTIR articles 88

 APPLICATION NOTE AN52335

Analysis of artists’ pigments by far-infrared

microspectroscopy
Authors Introduction
David W. Schiering , Ron Rubinovitz ,
1 2
The chemical identity of materials used in art pieces is of key importance to
Anthony W. Didomenico1, their authentication and preservation. By comparing the materials used relative
Beth Price3, and Kate Duffy3 to the time period in which the art was purportedly generated, conservators
1. Czitek, 6 Finance Dr., Danbury, CT 06810
can quickly distinguish authentic pieces from well-made copies. Furthermore,
2. Thermo Fisher Scientific, 4410 Lottsford Vista identification of the materials used can also guide conservators on the
Road, Lanham, MD, 20706
3. Philadelphia Museum of Art, PO Box 7646 proper cleaning and preservation techniques to slow the natural degradation
Philadelphia, PA 19101 processes of those art pieces. In many cases, it’s also important to identify
previously applied preservation materials.

Fourier transform infrared (FTIR) spectroscopy is a well-established technique

to analyze and identify artists’ materials.1-10 Measurements are typically
performed in the mid-IR spectral region (4000-400 cm-1). The KBr optics
commonly used in FTIR spectrometers impose a low frequency cutoff at
about 400 cm-1. Mercury-cadmium-telluride (MCT) detectors employed in
FTIR microscopes further restrict the low frequency cutoff to the 700-500 cm-1
range. These low frequency cutoffs limit the effectiveness of mid-IR in
identifying many inorganic pigments found in artwork, which often have
characteristic absorption bands in the far-IR region (600-10 cm-1). Raman
spectroscopy, another primary vibrational spectroscopy technique for the
molecular identification of cultural heritage materials, has a standard spectral
range well below 400 cm-1. However, fluorescence arising from both pigments
and binding media is an often-encountered problem in the analysis of artistic
pigments using Raman spectroscopy.11

The utility of far-IR spectroscopic characterization of pigment materials has

been demonstrated for macro samples.12-15 The far-IR microspectroscopy
analysis of pigments, however, has been an unexploited field due to such
challenges as the low radiance of thermal sources, relatively low responsivity
of room temperature detectors, and reduction of throughput by aperture
masks. Recent advances in optical materials and design have overcome many
of these challenges and opened the far-IR region for the microspectroscopic
application in pigment characterization. In this application note, we
Figure 1: Nicolet iS50 FTIR spectrometer with
the SurveyIR microspectroscopy accessory in demonstrate the microspectroscopic analysis of a number of pigments using
the sample compartment. the Thermo Scientific™ Nicolet™ iS50 FTIR Spectrometer and the Czitek™
SurveyIR™ Microspectroscopy Accessory (Figure 1).
Experimental Results and Discussion
Measurements were done in attenuated total reflectance To demonstrate the utility of the far-IR method and
(ATR) mode using the SurveyIR diamond ATR attachment, performance, mid- and far-IR spectra of a coated
or by reflectance with the samples on a low-E glass paperboard were collected. The coating on the
microscope slides. Spectra were collected using 256 paperboard contained calcium carbonate (CaCO3),
scans at 8 cm-1 spectral resolution. Mid-IR spectra kaolinite (Al2O3 2SiO2 2H2O), and a polyester binder. The
were measured using a Ge on KBr beamsplitter and coating spectra were collected in two modes using the
DLaTGS detector with a KBr window allowing collection SurveyIR diamond ATR microspectroscopy attachment.
from 4000 to 400 cm-1. Far-IR spectra in the range of First, a spectrum was collected using a conventional
1800-50 cm-1 were collected using a solid substrate KBr substrate beamsplitter and KBr window DLaTGS
beamsplitter and a DLaTGS detector with a polyethylene detector. This spectrum is displayed in Figure 2B in the
window. Polyethylene is commonly used as a window 1850-400 cm-1 region. The optical configuration of the FTIR
material in the far-IR, but its strong absorbances around was then changed to the far-IR mode. The IR spectrum
1460 cm-1 and 770 cm-1 do not allow measurements using this configuration was measured from 1850-50 cm-1
in these specific regions. Changing the beamsplitter and is displayed in Figure 2C. The overall spectral quality
between measurements was easily accomplished with and signal-to-noise ratios are quite satisfactory. Numerous
the automatic beamsplitter exchanger (ABX) on the absorption bands originating from calcium carbonate
Nicolet iS50 spectrometer. The unique ABX design allows and kaolinite are observed below 400 cm-1 outside of the
automated switching between measurement modes range of MCT detector and KBr window cut-offs. While
with no loss of purge. No modifications of the SurveyIR these inorganic materials also exhibit features in the mid-
accessory were necessary to allow it to measure spectra IR, extending the range into the far-IR region provides
in both mid- and far-IR regions. more spectral details that can be beneficial in identifying
materials or components in mixtures.
Figure 2: Measurement of coated paperboard
by ATR. (A) Image of sample area through
diamond ATR crystal before and after contact,
allowing the user to verify the sample area
being measured. (B) Mid-IR spectra with
characteristic low end cut-offs of MCT and
DLaTGS detectors indicated by dashed
lines. (C) Spectral features seen in the far-IR
spectrum indicative of calcium carbonate and
kaolinite seen below the KBr and MCT low
wavenumber cut-offs. Spectral regions around
1460 and 770 cm-1 are blanked due to the
strong absorbance of polyethylene detector
window at these locations.

The mid-IR and far-IR spectra of a modern paint are

shown in Figure 3, as another comparison. These
spectra were recorded using the SurveyIR diamond
ATR microspectroscopy attachment. The observed
mid-IR spectrum is consistent with polymers of ethyl
acrylate and methyl methacrylate,16-18 typical for a
modern polymer-based paint binder. However, the far-
IR spectrum allows the identification of the inorganic
pigment, zinc oxide, or “zinc white” since it contains a
characteristic absorbance peak at 380 cm-1. Clearly,
measurement of the mid-IR spectrum alone would not
allow for a positive identification of zinc white. Figure 3: White paint containing zinc oxide. Top: Mid-IR spectrum
showing only shoulder of zinc oxide absorption at 380 cm-1 (dashed
line). Bottom: Far-IR spectrum clearly showing zinc oxide’s absorption
peak at 380 cm-1.
Far-IR microspectroscopy was also employed in the indicated in the figure. The absorption bands between
analysis of pigments from a furniture piece in the collections the yellow pigment and the orpiment reference spectrum
of the Philadelphia Museum of Art (PMA). The piece, a match well and agree with previous reports.22-23 The
wardrobe shown in Figure 4, was purportedly fabricated in results unequivocally indicate that the yellow pigment
China in the 18th century during the Qing dynasty. To help is orpiment. In the cases of both pigments, the lack of
plan the conservation treatments, the decorative surfaces features in the mid-IR regions underscores the importance
were analyzed to determine whether they are consistent of far-IR measurements.
with the 18th century in China or are later additions. A cross-
section from a sample taken from an interior portion of the
wardrobe is also shown in Figure 4.

A B
Figure 5: Far-IR ATR spectra: (A) images of red pigment extracted from
Figure 4. Qing dynasty lacquer clothes wardrobe and inset shows a wardrobe sample, (B) far-IR ATR spectrum of wardrobe sample and (C)
cross-section from the interior used for analysis. (A) Exterior, (B) interior, vermillion reference.
and (C) cross-section analyzed from the compound wardrobe. The
asterisk represents the original decorative surface. The layers above
this point are later additions. Accession number: 1940-7-2, Philadelphia
Museum of Art; Artist/maker unknown Qing Dynasty (1644–1911);
Lacquered wood with painted and gilt decoration; brass fittings.
Purchased with the Bloomfield Moore Fund, 1940.

Far-IR microanalysis was performed on the red and

yellow layers of the cross-section. The images obtained
from the red pigment particles are shown in Figure 5,
Section A. Spectra of the red layer obtained by diamond
ATR microscopy and a vermilion reference are shown in
Figure 5, sections B and C, respectively. The vermilion
reference was from the Forbes pigment collection at
Harvard University.19 The two spectra exhibit a high degree
Figure 6: Far-IR ATR spectra: (A) images of yellow pigment extracted
of similarity and confirm a positive identification. Vermilion
from the wardrobe sample, (B) far-IR reflectance spectrum of yellow
is produced from the mineral cinnabar, HgS, mercury (II) wardrobe sample and (C) orpiment reference.
sulfide. The observed bands are consistent with those
from trigonal α-mercuric sulfide (cinnabar mineral) and are Conclusions
attributed to the lattice vibrations of the crystal.20 Vermilion Mid-IR FTIR spectroscopy is an effective analytical tool
was a common red pigment used in the production of for artworks, and the combination with microscopy is
Chinese lacquerware21 and its presence provides critical particularly beneficial as minimal sample is required for
evidence of the wardrobe’s authenticity. The captured analysis. However, many materials, particularly pigments
images of the yellow particles (A), the far-IR spectra of the containing inorganic components, cannot be fully
yellow lacquer layer pigment (B), and orpiment reference characterized with FTIR microscopes configured only
(As2S3) (C) are shown in Figure 6. Both spectra were for mid-IR operation. The analyses of zinc white and two
recorded in the external reflectance mode from samples layers of an antique artifact demonstrate the benefits of
flattened onto low-E glass substrates. An aperture mask combining far-IR spectroscopy with microspectroscopy
corresponding to 250 microns at the sample was used accessories, which allow for more insightful
to target the pigment extracted from the yellow layer. The characterization of pigment materials. The chemical
reference orpiment pigment was also from the Forbes information gleaned from the interpretation of these
collection. Orpiment is a mineral produced geothermally spectra can be used to establish authenticity through
and was used as a yellow pigment from antiquity into dating of artworks, characterize previous conservation
the early 20th century. Prominent absorption bands are attempts, and plan new conservation campaigns.
References
1. Infrared Spectroscopy in Conservation Science, Michele R. Derrick, Dusan Stulik, and James M. Landry, J. Paul
Getty Trust, Getty Conservation Institute, Los Angeles, 1999.

2. Meiluna, Raymond J., James G. Bentson, and Arthur Steinberg. Analysis of Aged Paint Binders by FTIR
Spectroscopy. Studies in Conservation 35 (1990).

3. Sloggett, Robyn, Caroline Kyi, Nicole Tse, Mark J. Tobin, Ljiljana Puskar, and Stephen P. Best. Microanalysis of
Artworks: IR Microspectroscopy of Paint Cross-sections. Vibrational Spectroscopy 53.1 (2010): 77-82.

4. Newman, Richard. Some Applications of Infrared Spectroscopy in the Examination of Paintings Materials. Journal
of the American Institute for Conservation 19.1 (1979): 42-62.

5. Manzano, E., N. Navas, R. Checamoreno, L. Rodriguez-Simon, and L.F. Capitan-Vallvey. Preliminary Study of UV
Ageing Process of Proteinaceous Paint Binder by FTIR and Principal Component Analysis. Talanta 77.5 (2009):
1724-731.

6. Jonsson, Julia, and Tom Learner. Separation of Acrylic Paint Components and Their Identification with FTIR
Spectroscopy. Proceedings of the Sixth Infrared and Raman Users Group Conference (IRUG6). Florence, Italy
March 29th-April 1st 2004. Ed. Marcello Picollo. Florence: Il Prato, 2005. 58-68.

7. Hodson, J., and J.A. Lander. The Analysis of Cured Paint Media and a Study of the Weathering of Alkyd Paints by
Fourier Transform Infra-red/photoacoustic Spectroscopy. Polymer 28.2 (1987): 251-56.

8. Lang, Patricia L., Chad D. Keefer, Jessica C. Juenemann, Khoa V. Tran, Scott M. Peters, Nancy M. Huth, and
Alain G. Joyaux. The Infrared Microspectroscopic and Energy Dispersive X-ray Analysis of Paints Removed from a
Painted, Medieval Sculpture of Saint Wolfgang. Microchemical Journal 74.1 (2003): 33-46.

9. Olin, J.S. The Use of Infrared Spectrophotometry in the Examination of Paintings and Ancient Artifacts. Instrument
News 17 (1966): 1.

10. McClure, A., J. Thomson, and J. Tannahill. Infrared Spectra of Ninety-six Organic Pigments. Journal of Oil and
Colour Chemists’ Association 51 (1968)

11. López-Gil, Ruiz-Moreno, Miralles. Optimum acquisition of Raman spectra in pigment analysis with IR laser diode
and plulsed UV irradiation Journal of Raman Spectroscopy. 37 (2006); 966-973

12. Kendix, Elsebeth L., Silvia Prati, Edith Joseph, Giorgia Sciutto, and Rocco Mazzeo. ATR and Transmission
Analysis of Pigments by Means of Far Infrared Spectroscopy. Analytical and Bioanalytical Chemistry 394 (2009):
1023-032.

13. http://lisa.chem.ut.ee/IR_spectra/

14. Afremow, Leonard C., and John T. Vanderberg. High Resolution Spectra of Inorganic Pigments and Extenders in
the Mid-Infrared Region from 1500 cm-1 to 200 cm-1. Journal of Paint Technology 38.495 (1966): 169-202

15. C. Karr and J. Kovach. Far-Infrared Spectroscopy of Minerals and Inorganics. Appl. Spectrosc. 23, 219 (1969).

16. M.R. Nelson, Authentic or Not, ChemMatters, April 2011, pp. 15-17.

Find out more at

thermofisher.com/is50 and thermofisher.com/surveyIR
©2018 Thermo Fisher Scientific Inc. All trademarks are the property of Thermo Fisher Scientific and its subsidiaries unless
otherwise specified. AN52335_E 02/18M
 APPLICATION NOTE AN52813

Attenuated total reflection FTIR imaging

of “soft chew” formulations
Author Introduction
Ron Rubinovitz, Ph.D. In recent years, the pharmaceutical
Thermo Fisher Scientific, USA formulation type in which the active
ingredients are dispersed in a “soft
Keywords chew” matrix has become increasingly
ATR, FTIR, imaging, microscopy, popular for a number of over-the-
formulations counter (OTC) dosage forms. The
sticky and oily nature of these dosage
Thermo Fisher Scientific forms, however, poses a particular Figure 1: Imaging ATR accessory (left) and
solutions challenge in sample handling for Nicolet iN10 MX infrared microscope (right)

Nicolet iN10 MX Infrared Imaging FTIR measurements. For example, it

Microscope is often difficult to cut these samples sufficiently thin for transmission analysis,
while simply pressing them flat would destroy the innate component distribution
within the samples. Reflectance spectra, on the other hand, tend to be of poor
quality. The Attenuated Total Reflectance (ATR) measurement of a sample cross
section could offer a potential solution since this technique would expect to
yield high quality spectra regardless of sample thickness. The “stickiness” of
these samples, however, leads to “sample carry over” and complicates single
point ATR mapping. In addition, maps acquired by ATR typically require longer
measurement times due to the time required to raise and lower the sample stage
at each measurement point.

In this note, we demonstrate the use of an imaging ATR accessory that’s well
suited for the analysis of these samples. The accessory makes contact with the
sample just once, while still permitting measurements across the sample surface.
Additionally, germanium ATR microscope measurements offer the benefit of a
four times magnification due to presence of germanium (refractive index n=4)
instead of air (n=1) at the sample interface, resulting in enhanced spatial resolution
compared to non-ATR measurements.

Experimental
Spectra were measured using a Nicolet iN10 MX infrared imaging microscope
(Figure 1), configured with the following detectors: a room temperature DTGS, a
single point MCT-A and a linear array MCT-A detector. Each sample cross section
was cut to a size of approximately 5 x 5 x 2 mm (l x w x h) and was placed on the
sample post of an imaging ATR microscope accessory (Figure 1) and raised to
make contact with the germanium ATR crystal for measurement.
Unless otherwise indicated, spectra were collected using
the array detector in “ultra-fast” mode where a single scan
per step at 16 cm-1 resolution was collected at each map
point at a spatial resolution of 6.25 microns. Sample and
background maps of about 625 x 625 microns consisting of
about 11,000 spectra were measured in approximately 1.5
minutes each.

Results
Antacid product component distribution
Representative map spectra and their comparison to the Figure 2: Maps showing distribution of main components of
library spectra of pure materials are shown in Figure 2. antacid product (top) and representative spectra shown in red,
Examination of individual map spectra showed that compared to library spectra shown in blue (bottom)

each spectrum was dominated by features of a specific

sample component, although weaker features of the
other major components can also be seen. Maps of
the active material (calcium carbonate), and two of the
main non-active components, oil and corn syrup based
on baseline corrected peak heights at 872, 1746, and
1020 cm-1, respectively are shown in Figure 2. It is clear
from the similarities in the active component and corn
syrup maps that the calcium carbonate is found primarily
in the corn syrup regions. Figure 3: Comparison of maps of active component measured using
an array (A) and single point detector (B)
In Figure 3, the quality of the maps created from array
spectra is verified by comparing two maps of the same
sample measured with array and single point MCT-A
detectors, respectively. Each map shows the intensity
of the calcium carbonate peak at 872 cm-1 and contains
approximately 2000 spectra. The array map of single
scan spectra at 16 cm-1 collected in just 18 seconds
shows the same features as the single point detector
map collected at 8 cm-1 resolution and 16 scans per
spectrum that required about two hours to collect.

Joint health supplement component distribution

Due to the low spectral intensity of the glucosamine Figure 4: Maps of components obtained with MCR Analysis (top)
in this product, a multivariate curve resolution (MCR) and MCR components shown in red, compared to library spectra
analysis was performed which readily extracted the shown in blue (bottom)

spectra of glucosamine and, as expected, the more

obvious oil and corn syrup features. The resulting maps
are shown in Figure 4, along with the comparison of the
MCR components and pure spectra. The minor presence
of glucosamine in the measured spectra is demonstrated
in Figure 5, using principal component reconstructed
spectra. A map spectrum of mostly corn syrup is
subtracted from a map spectrum that MCR analysis
indicated contained both glucosamine and corn syrup,
resulting in spectral features that compare well to the Figure 5: Spectral subtraction confirming the
reference spectrum of glucosamine. presence of glucosamine as found by MCR analysis
Conclusions
In this note, the measurement of the key component
distribution in two “soft chew” products by ATR FTIR
imaging is described. Maps of sample components with
relatively intense features were created by straightforward
peak-height methods, while significantly weaker features
were readily mapped by using MCR techniques. For each
sample, the maps were collected in just a few minutes.
The improvement in analysis speed was accomplished
through the use of an array detector and a “fit-for-purpose”
accessory that eliminates the need to raise and lower
the microscope stage at each measurement point. The
described methodology should have general applicability
for QA/QC and analytical laboratories in pharmaceutical,
food and associated industries performing similar analyses.

Find out more at thermofisher.com/ftir

©2018 Thermo Fisher Scientific Inc. All trademarks are the property of Thermo Fisher Scientific and its
subsidiaries unless otherwise specified. AN52813_E 05/18M
APPLICATION NOTE AN52985

Protein secondary structure elucidation using FTIR spectroscopy

Author Abstract
Suja Sukumaran Fourier-transform infrared (FTIR) spectroscopy is one of the most
Thermo Fisher Scientific, USA versatile analytical tools used across various disciplines. In this study, the
Thermo Scientific™ Nicolet™ iS10 and Nicolet iS50 FTIR spectrometers,
Keywords equipped with Attenuated Total Reflection (ATR)-FTIR and Transmission-
FTIR, ATR, protein structure FTIR, were used for the determination of protein secondary structures.
elucidation, Biocell calcium Structure calculations based on a protein database as well as spectral
fluoride cell, ConcentrateIR2 ATR, deconvolution are discussed. The analyses were quick and easy.
transmission
Introduction
Protein secondary structure describes the repetitive conformations of
proteins and peptides. There are two major forms of secondary structure,
the α-helix and β-sheet, so named by the patterns of hydrogen bonds
between amine hydrogen and carbonyl oxygen atoms that create the
peptide backbone of a protein.1 Understanding protein secondary structure is
important to gain insight into protein conformation and stability. For example,
temperature dependent analysis of the secondary structure is critical in
determining storage conditions for maintaining active therapeutic proteins.2
Protein secondary structure is also crucial in understanding the structure-
function relationship and enzyme kinetics of various proteins.3

FTIR has long been established as a powerful analytical technique to investigate

protein secondary structure and local conformational changes.1, 4 A typical
protein infrared (IR) spectrum often contains nine amide bands, with vibrational
contributions from both protein backbone and amino acid side chains. Among
which, of particular pertinence to protein secondary structure are Amide I and
Amide II bands. The absorptions associated with C=O stretching are denoted
as Amide I, whereas those associated with N—H bending are Amide II. Since
both C=O and N—H bonds are involved in the hydrogen bonding between
different moieties of secondary structure, the positions of both Amide I and
Amide II bands are sensitive to the secondary structure composition of a
protein,3, 4 although the Amide II band is widely viewed as a less useful predictor
for quantifying the secondary structure of proteins.
The shifts in the Amide I band are often small compared to the two CaF2 windows. CaF2 windows are suited for
the intrinsic width of the band, resulting in one broad peak water-based sample analysis. As water has a significant
instead of a series of resolved peaks for each type of the absorption peak at 1645 cm-1 region, a small path length of
secondary structure. Mathematical procedures such as 6 μm can effectively avoid saturated water peaks.
Fourier self-deconvolution and second derivatives can be
used to resolve the overlapping bands for the quantitative A purged Nicolet iS10 FTIR spectrometer, equipped
analysis of protein secondary structure.3 Table 1 shows with a DTGS detector, was used for transmission
the secondary structure band assignments for proteins in analysis. The scan parameters used were 256 scans,
water. Note that all assignments are depicted as a range, resolution of 4 cm-1. The Thermo Scientific™ Smart
as the exact position of each peak varies from protein OMNI-Sampler™ Transmission Accessory allows for a
to protein due to the differences in hydrogen bonding quick purge of the chamber, eliminating the need for
interactions and the environment of the proteins. water vapor subtraction in most analyses. Secondary
structure analysis of the buffer-subtracted spectra was
Band Assignment carried out using the built-in feature of the PROTA-3S™
Secondary Structure
in Water FT-IR Protein Structure Analysis Software. Secondary
structure calculation in PROTA-3S software is based
α-Helix 1648-1657 cm-1 on a database of 47 secondary structures (for more
β-Sheet 1623-1641 cm-1 information visit www.btools.com).
(high frequency component) 1674-1695 cm-1
For ATR analysis, a ConcentratIR2™ Multiple Reflection
Random 1642-1657 cm-1
ATR Accessory (Harrick Scientific Products, Inc.
Coils 1662-1686 cm-1 Pleasantville, NY) with diamond crystal was used in
a Nicolet iS50 FTIR spectrometer equipped with an
Table 1: Secondary structure band assignments for protein in water.2
mercuric cadmium telluride (MCT) detector. The diamond
With a range of sampling techniques, including ATR has ten internal reflections with a nominal angle of
transmission, ATR, and infrared reflection absorption incidence of 45°. A 10 μL protein solution in phosphate
spectroscopy (IRRAS), FTIR is particularly advantageous in buffer was dried on the surface of the ATR crystal under
terms of its versatility and general applicability compared a stream of nitrogen. Scan parameters used were 256
to other analytical techniques for protein secondary scans and a resolution of 4 cm-1. Secondary structure
structure analysis. Protein sample forms suitable for FTIR determination was carried out using the peak resolve
analysis include lyophilized powders, water solution, and feature of the OMNIC software.
colloids, to name a few. We report herein two examples
of protein secondary structure determination using Results and discussion
transmission-FTIR and ATR, respectively. Both methods Transmission-FTIR with Biocell
are fast, consume a minute amount of sample, and require Figure 1 shows the overlay of three FTIR spectra:
minimal sample preparation. phosphate buffer, cytochrome C at 6 mg/mL and
12 mg/mL in phosphate buffer, respectively. At first
Experimental glance, the spectra are predominantly water bands. The
All proteins were procured from Sigma-Aldrich (MO, USA) three spectra show little difference, even at a high protein
and used as received. For the transmission studies, a concentration of 12 mg/mL.
BioCell™ Calcium Fluoride Cell (Biotools,
Jupiter, FL) was used, and all
measurements were carried out at
ambient temperature. A 10 μL protein Figure 1: Transmission-FTIR spectra for
cytochrome C in phosphate buffer (cytc_12)
solution was placed at the center of at 12 mg/mL and 6 mg/mL (cytc_6), and
the window, and the protein solution phosphate buffer blank.
was sandwiched between the two CaF2
windows, and placed in the holder.
The concentration of protein tested
was between 6 and 12 mg/mL. A 6 μm
path length was created by sandwiching
Next, the buffer spectrum was subtracted from the raw
protein spectra using the PROTA-3S software, and the
results are shown Figures 2A (cytochrome C) and 2B
(concanavalin). The Amide I and II peaks are clearly
discernible in both spectra. The Amide I peak position
for cytochrome C spectra is 1654 cm-1, suggesting an
α-helix dominant secondary structure. For Concanavalin
A, the Amide I peak centers at 1633 cm-1, and there is
also a noticeable shoulder peak at 1690 cm-1 (red circle),
indicative of the β-sheet component and its associated
high-frequency component.2

Table 2 summarizes the secondary structure prediction

using the PROTA-3S software. The cytochrome C has
45% α-helix and 5% β-sheet, whereas Concanavalin A has
42% β-sheet and 4% α-helix. Differences in secondary
structure composition between X-ray and FTIR data are
likely due to the physicochemical state of the protein
samples such as crystalline versus solution, temperature, Figure 2: FTIR spectra of (a) cytochrome C and (b) Concanavalin A
pH, buffer conditions, etc. Furthermore, different after the buffer spectrum was subtracted using PROTA-3S software.

prediction algorithms could have slightly varying outputs.7

Notwithstanding the differences in analytical technique, α-Helix (%) β-Sheet (%) Random (%)
sample state, and prediction algorithms, the secondary
Protein FTIR X-ray FTIR X-ray FTIR X-ray
structure elucidation by FTIR using PROTA-3S software
is largely in line with that from X-ray. Transmission-FTIR Cytochrome C 45 41 5 0 50 59
measurements combined with PROTA-3S software offer a
Concanavalin A 4 0 42 48 54 52
facile and fast means to analyze the secondary structure
of proteins in solution2, 3 with minimal sample prep.
Table 2: Comparison of secondary structure calculation from FTIR
(PROTA-3S) and X-ray data.
ATR-FTIR with ConcentrateIR Accessory
When the quantity and concentration of protein are Only those buffers with minimum or no peaks in the
limited, FTIR measurements with the ConcentratIR2 ATR Amide I and II region should be selected. Figure 3 shows
accessory offer a better alternative than transmission-FTIR the ATR-FTIR spectra of BSA in phosphate buffer, dried
spectroscopy. The unique design of this ATR accessory on the crystal from a 1 mg/mL solution. In addition to
allows for the direct measurement of protein powders, gels, the Amide I and II bands, there are spectral features
solutions as well as proteins dried on the ATR surface. of the side chain, such as 1515 cm-1 from tyrosine
When concentrating proteins on the crystal surface, caution and 1498 cm-1 from aspartic acid. Side chain peaks
should be exercised in buffer selection since buffer will also are critical in the elucidation of protonation and de-
concentrate on the surface of the crystal. protonation states of various amino acids.2

Figure 3: Amide I and II for 1 mg/mL BSA

analyzed using ConcentratorIR2 ATR on the
Nicolet iS50 FTIR spectrometer equipped
with an MCT detector.
Peak deconvolution of the Amide I peak (Figure 4) of Conclusion
BSA was carried out using the OMNIC software. It is In this note, we have demonstrated two examples of protein
important to note that second derivative analysis is often secondary structure elucidation using FTIR spectroscopy.
performed prior to deconvolution to clearly identify the Transmission-FTIR measurements combined with PROTA-
peaks required for peak fitting.2 In the current study, the 3S software provides a facile means to analyze secondary
second derivative peaks obtained (results not shown) well structure of proteins in solution with minimal sample
correlate to the secondary structure peak assignments preparation. When the quantity and concentration of protein
in Table 1. In order to obtain a good peak shape for are limited, ATR-FTIR offers a better alternative by drying
peak fitting, a baseline correction on the Amide I region the proteins in ATR crystals directly.
was also performed. Baseline correction also effectively
excluded the contributions from the Amide II region. References
The deconvolution of Amide I resulted in 5 peaks, and 1. Elliott, A., Ambrose, E. J. Structure of synthetic
the area under each peak was then evaluated against polypeptides, Nature (1950) 165, 921-922.
the total area. Amide I peak deconvolution shows 2. Jackson, M., Mantsch, H.H. The use and misuse of
a secondary structure composition of 47% α-helix, FTIR spectroscopy in the determination of protein
3% β-sheet, 24% coils, and 26% random, similar to structure, Crit. Rev. Biochem. Mol. Biol. (1995) 30,
published FTIR5 and X-ray data. 95-120.
3. Barth, A. Infrared spectroscopy of proteins, Biochim.
Biophys. Acta (2007) 1767, 1073-1101.
4. Byler, D.M., Susi, H. Examination of the secondary
structure of proteins by deconvolved FTIR spectra,
Biopolymers (1986) 25, 469-487.
5. Surewicz, W.K., Mantsch, H.H. New insight into
protein secondary structure from resolution-enhanced
infrared spectra, Biochim. Biophys. Acta (1988) 952,
115-130.
6. Sukumaran, S., Hauser, K., Maier, E., Benz, R.,
Mantele, W. Tracking the unfolding and refolding
pathways of outer membrane protein porin from
Figure 4: Peak deconvolution of Amide I peak of BSA using Peak
Resolve function of OMNIC software. Paracoccus denitrificans, Biochemistry (2006) 45,
3972–3980.
7. Klose, D., Janes R.W. 2Struc – the protein secondary
structure analysis server, Biophysical Journal (2010)
98, 454-455.

Find out more at www.thermofisher.com/FTIR

©2017 Thermo Fisher Scientific Inc. BioCell and PROTA-3S are trademarks of BioTools, Inc. ConcentratIR2
is a trademark of Harrick Scientific Products, Inc. All other trademarks are the property of Thermo Fisher
Scientific and its subsidiaries unless otherwise specified. AN52985_E 10/17
 APPLICATION NOTE AN52307

A case study of using FT-NIR for pharmaceutical hot

melt extrusion process monitoring
Author Introduction
Herman He , Scott Martin , Anh Vo ,
1 1 2
Continuous manufacturing, such as hot melt extrusion (HME), is gaining
Michael Repka2 and Rui Chen1 increasing interest in the pharmaceutical industry due to its high manufacturing
Thermo Fisher Scientific, USA
1
efficiency and economic benefits.1 HME is primarily a mechanical/shear
2
School of Pharmacy, University of Mississippi mixing process, but can also include a chemical reaction process in some
formulations. During HME an active pharmaceutical ingredient (API) is
Keywords melted, mixed, and bonded with thermoplastic polymers and/or other
pharmaceutical extrusion, hot binding materials to form a unique polymeric structure. Judiciously designed
melt extrusion (HME), process polymers can promote desired API/polymer bonding structure, improve drug
analytical technology (PAT), bio-availability, and achieve unique drug delivery profiles.2, 3, 4 HME process
FT-NIR Spectroscopy, active operation parameters, such as screw design and rotational speed, zone
pharmaceutical ingredient (API), temperature, and residence time must be controlled precisely to ensure the
partial least square (PLS), principle consistency and the product quality. In the new era of pharmaceutical process
component analysis (PCA) development, quality by design (QbD), one of the key initiatives from the
FDA,5, 6 requires a thorough understanding of all critical process parameters
Application benefits so that an effective process control strategy can be implemented for more
• Real time in-process monitoring uniform product quality and more efficient production operations.

• Nondestructive and fast

As an invaluable in-process analytical tool, Fourier transform near-infrared
• Provides chemical structural (FT-NIR) spectroscopy in conjunction with statistical regression techniques has
information been used to provide near real-time chemical information. FT-NIR spectroscopy
has been widely used as a process analytical technology (PAT) and QbD tool
Thermo Fisher Scientific for process development and manufacturing in pharmaceutical industry to
solutions monitor and/or control such processes as drying, blending, and extrusion.7, 8 It
• Antaris FT-NIR Analyzer is a simple, fast, and nondestructive technique that provides multi-constituent
analysis of virtually any matrix without the need for sample preparation
• Process 11 Twin Screw Extruder
and manipulation. More importantly, NIR provides molecular level chemical
• Result software information through a direct measurement of the material being processed. It is
sensitive to the API concentration in a blend, and the interactions between the
• TQ Analyst software
API and its binding polymers.

In this application note, we demonstrate the use of FT-NIR for in-line API
concentration monitoring during an HME process. The considerations in deriving
an appropriate partial least square (PLS) model are discussed. In addition, the
feasibility of employing NIR spectral responses as an indicator for HME process
stability is also explored.
Experimental
An 11-mm twin-screw extruder, Thermo Scientific™ Pre-blended ketoprofen (API) and Eudragit® L100-55
Process 11 (Thermo Fisher Scientific, Karlsruhe, (polymer), with their chemical structures shown in Figure
Germany) was used for the HME processes. Figure 1 2, were fed into the extruder at Zone 1 position. Zones 2
is a schematic showing the screw profile design as well through 8 were mixing, conveying and discharge zones.
as process configuration, and the effective length-to- Zone 9 is a die that reduces the barrel configuration to a
diameter ratio (L/D) for the system was defined as 40:1. single circular orifice.
Three mixing zones were configured for the process.
polymer/API
pre-blend

Figure 1: Schematic of a Thermo Scientific Process 11 Twin Screw Extruder configured in this study.

Figure 2: Chemical structures of ketoprofen and Eudragit® L100-55.

A Thermo Scientific™ Antaris™ II MDS FT-NIR The target blending ratio of ketoprofen and Eudragit
spectrometer (Thermo Fisher Scientific, Madison, L100-55 was 50:50 (w/w). The temperature in the mixing
Wisconsin), equipped with a HME diffuse reflectance and melting zones was set at 120 °C. The material feed
probe, was used for spectral acquisition. The reflectance rate was 100 g/h. NIR data acquisition was performed
probe has standard 1/2-20 UNF mounting threads and using the Thermo Scientific™ RESULT™ v3 software
was screwed into one of the sensor ports of the exit package. The spectral range was 4000–10,000 cm-1
block, between the exhaust gas vent and the exit orifice. and the resolution was 8 cm-1. The number of spectral
Since the hot melt is a translucent liquid at the sampling scans averaged was 16, or 8 seconds per spectrum. A
location, a metal reflector was placed on the opposite background reference spectrum was acquired using the
side of the probe to enhance the NIR signal. The gap transmission sampling module at the beginning of each
between the reflector and the probe surface was ~2 mm. set of experiments.
The exit block was heated at 120 °C by an electric heater
controlled by the extruder PLC.
Results and discussion To build a calibration model to predict the API
Building a PLS model to monitor API concentration during HME, the pre-blend API/polymer
concentration for HME ratio was varied from 40 to 60%, bracketing the
Figure 3A shows the overlaid NIR spectra of ketoprofen, target API concentration of 50% (w/w). The normal
Eudragit, and a 50:50 (w/w) extrudate. Both ketoprofen process temperature was set at 120 °C, but +/- 10 °C
and Eudragit are powders at room temperature, and their disturbances were introduced to simulate possible
absorption spectra were measured using an integration temperature variation in manufacturing process. The
sphere module of the Antaris II spectrometer. The feed rate of 100 g/h remained constant throughout
spectrum of the extrudate was acquired with an NIR the experiment. After each API concentration change
HME probe during an extrusion process. All spectra have was introduced, adequate time was allowed to ensure
strong features in the first overtone C-H stretching region a steady state was reached before the next process
(~6000 cm-1). In addition, both Eudragit and the extrudate change. The spectral data collected during steady states,
have noticeable peak features in the first overtone O-H including those at different temperatures, were used
stretching region (~6900 cm-1). as calibration samples. NIR spectra were collected at
a 20-second interval. The calibration sample set has a
total of 85 spectra at 7 concentration levels: 40%, 45%,
47.5%, 50%, 52.5%, 55%, and 60%.
A
Figure 3B shows the overlaid NIR spectra of the
calibration samples. There is a noticeable baseline up-
drift in all spectra. While it is difficult to ascertain the
exact causes for this up-drift, one likely contributing
factor is the scattering by the air bubbles formed
during the HME process and the variation in effective
pathlength. Because of the relatively low material feed
rate (100 g/h); hence, high air/material ratio, the vent
nozzle was not able to completely remove air and
moisture from the hot melt, resulting in micro-size air and
water vapor bubbles.
B
The Norris second derivative10 was first applied to the
raw spectra to remove the baseline drift, followed by a
standard normal variant (SNV) to minimize the spectral
pathlength variation. A PLS model11 to estimate the
ketoprofen concentration was then derived using the
Thermo Scientific™ TQ Analyst™ software. All spectral
data were mean-centered in the PLS regression, and
the spectral range of 5500 to 6650 cm-1 was used
for the correlation. A total of 85 spectra were divided
Figure 3. (A) Absorption NIR spectra of ketoprofen, Eudragit L100-
55, and a 50/50 (w/w) ketoprofen/Eudragit extrudate in full scale. into two groups: calibration (73 spectra) and validation
(B) Transflectance NIR spectra in common scale acquired during (12 spectra). PLS finds its significant model factors
extrusion processes.
from the Leave-One-Out cross validation. The number
NIR spectra are often complex and contain broad of significant PLS factors represents the number of
overlapping absorption bands. Chemical, physical, independent variables that affect sample spectral
and structural properties of all species present in a responses, such as concentration, impurities, density,
sample influence the measured spectra. As a result, the opaqueness, or sample color. In the current case,
measured NIR spectra depend on more than one variable the developed PLS model (Figure 4) has a correlation
simultaneously; hence, multivariate. Chemometrics uses coefficient of 0.998 and a root mean squared error of
mathematical and statistical procedures, such as PLS calibration (RMSEC) of 0.43%. The root mean squared
and principle component analysis (PCA), for multivariate error of prediction (RMSEP) from the validation sample
data analysis to filter information that correlates to a group is 0.62%.
certain property from a large set of data.
The calibration model was then applied to the spectra
collected during the HME process, and the results
are shown in Figure 5. As can be seen, the predicted
API% tracks the pre-blend API profile. With a feed
rate of 100 g/h, the process has a residence time of
~10 min, but it takes 2-3 multiples of the residence time
(approximately 20-30 min) to establish a new steady state
after the pre-blend ratio is introduced. Figure 4. Calibration results of API measurement.

Extracting the relevant portion of the multivariate NIR

spectral data is critical for a successful chemometrics
model. For example, process temperature has a profound
impact on measured NIR spectra because temperature
can affect samples both physically (e.g. sample density)
and chemically (e.g. the degree of hydrogen bonding).
Figure 6A shows the 2nd derivative NIR spectra of hot
melt samples collected at three process temperatures.
There are prominent spectral changes around 6900
cm-1, likely due to the increased H-O bonding at elevated
Figure 5. Overlay of the pre-blend profile and the API% trend based
temperatures. However, this spectral region was purposely on the PLS model.
excluded in the current PLS model so that the model can
keenly reflect the API concentration change without the
interference from temperature induced spectral variation.
The spectral range of 5500 to 6650 cm-1 was used instead
for the correlation. Figure 6B shows the API% change
with temperature ramping from 110 to 130 °C, while
maintaining a pre-blend API ratio of 50% (w/w) and feed
rate of 100 g/h, simulating potential temperature variations
in a real HME process. The predicted API% hovers around
49-51% throughout the experiment. The results suggest
that the developed PLS model for API% prediction is
robust with a +/- 10 °C temperature variation.

Process stability monitoring with PCA

Continuous manufacturing requires the production
process to be controlled in its steady state to ensure
smooth production and stable product quality. In an HME
process, the controllable process variables include the
speed of material flow, extrusion temperature distribution
profile, and screw speed, etc. Traditionally, an HME
process is considered at its steady state when operational
parameters such as torque, material temperature, and
pressure reach their equilibria. NIR, on the other hand,
not only provides chemical information of the samples,
including blending ratio and interactions between the
API and the binding material, but also is sensitive to
many physical properties of the sample, such as color,
temperature, and density. NIR spectra are therefore a
Figure 6. Process temperature robustness study: (A) temperature
more incisive, responsive, and comprehensive process impact on NIR spectra; and (B) prediction of API% with a +/- 10 °C
indicator for product quality control. temperature variation.
To simulate potential process fluctuations during a Similar to the temperature experiment, it takes about 2
manufacturing process, two process variables, feed minutes for the PC data points to shift to a new location in
rate and temperature, were introduced as process response to the feed rate change.
disturbances. Sample spectra were acquired at a
20-second interval. The raw spectral data was processed
with the 2nd derivative and SNV to remove spectral
baseline drift and variations in spectral path length, then
analyzed using PCA.9 The spectral wavelength region
used for the PCA calculation was from 9150 to 4710 cm-1.

PCA defines its principal component (PC) domain space

through an iteration of information extraction from a
calibration sample set, and converts each sample
spectrum consisting of thousands of data points into a
reduced, e.g. 2-3 dimensional, data point in a PC domain.
The number of dimensions in a PC domain is related to the
number of process variables that affect sample spectra,
such as concentration, degree of cross-link, sample
color, and density. Generally, the first several principal
components cover the majority of the spectral information.
Scores, or coefficients, of principal components are
projections of a measured sample spectrum in a PC
domain. By converting a sample spectrum into a PC
point, sample spectral responses during an HME process
can be visually displayed in a 2- or 3-dimensional trend
plot. Because NIR spectra are a direct measurement
Figure 7. PC score plot of (A) temperature response; and (B) feed
of the extrudate product, the trajectory of their PC rate response.
coordinates movement in a PC domain should be a better *Note: The API concentrations for the temperature and feed rate
experiments are different; hence, the PC1 / PC2 ranges are different in
representation of a HME process state. The trajectory
the plots.
from one cluster to another offers insight into the process
dynamics so that a more comprehensive control strategy Conclusions
can be developed to make the HME process more stable. Continuous manufacturing requires in-process
monitoring of critical process parameters to ensure
Figure 7A is a PC score plot of the temperature product consistency. In this note, FT-NIR in conjunction
experiment. The feed rate was maintained at 100 g/h. with chemometrics modelling for in-line HME process
The process temperature started at 120 °C. The sample monitoring is demonstrated. In the first case, a PLS
spectra form a tight cluster of PC points (middle) in the model was successfully developed for API concentration
PC domain. Once the process was stabilized, a 10 °C prediction by carefully selecting the relevant NIR spectral
temperature decrease was introduced in the temperature region. The developed model is also proven robust with
controller. The spectral response from the hot melts a +/- 10 °C temperature fluctuation. The second case is
changed accordingly. As a result, the resulting PC data concerned with the use of a PCA based model to monitor
point moves away from the previous cluster. It took about the process state shift in response to disturbances such
6 sampling times, or about 2 minutes for the sample PC as temperature and material feed rate. Because the
point to move from the 120 °C cluster to a new location NIR spectrum is a direct measurement of the extrudate
and start to form a new cluster. The distance between product, the trajectory of the PC movement offers insight
the two clusters represents some property shift in the hot to the process dynamics. Moreover, since the PCA method
melt. Similar PC point movement was observed for the relies solely on an accumulation of multiple batches of
temperature change from 110 to 130 °C. good process spectral data to define the steady state, it
does not require a calibration model. It is independent of
Figure 7B is a PC score plot of the feed rate response the product formulation and is easy to implement. Finally,
experiment. The temperature was maintained at 120 °C. because NIR is nondestructive and fast, an NIR-based
The process started with a 100 g/h material feed rate. Once quality monitoring methodology can be easily transferred
it was stabilized, the feed rate was increased to 120 g/h. from process development to manufacturing.
References
1. Schaber S.D., Gerogiorgis D.I., Ramachandran R., Evans J.M.B., Barton P.I. and Trout B.L. Economic analysis of
integrated continuous and batch pharmaceutical manufacturing: a case study, Ind. Eng, Chem. Res., 2011, 50:17,
10083-10092. doi: 10.1021/ie2006752.

2. Thiry J., Krier F., Evrard B. A review of pharmaceutical extrusion: Critical process parameters and scaling-up, Int.
J. Pharmaceut., 2015, 479:1, 227-240.

3. Breitenbach J. Melt extrusion: from process to drug delivery technology, Eur. J. Pharm. Biopharm., 2002, 54:2,
107-117.

4. Crowley M.M., Zhang F., Repka M.A., Thumma S., Upadhye S.B., Battu S.K., McGinity J.W., and Martin C.
Pharmaceutical applications of hot-melt extrusion: Part I, Drug Dev. Ind. Pharm., 2007, 33:9, 909-926.

5. FDA Guidance for Industry. PAT – a framework for innovative pharmaceutical development, manufacturing, and
quality assurance. Rockville, MD: Food and Drug Administration, 2004.

6. ICH Q8(R2), Pharmaceutical Development, Part I: Pharmaceutical development, and Part II: Annex to
pharmaceutical development, 2009.

7. Markl D., Wahl P.R., Menezes J.C., Koller D.M., Kavsek B., Francois K., Roblegg E., and Khinast J.G. Supervisory
control system for monitoring a pharmaceutical hot melt extrusion process, AAPS PharmSciTech, 2013, 14:3,
1034-44.

8. MacPhail N., Meyer R.F, Phillips, J.X., Gendron C.M., Smith-Goettler B. NIR Monitoring of a hot-melt extrusion
process, Spectroscopy Special Issues, 2011, 26:8.

9. Abdi H. and Williams L.J., Principal component analysis, Wires. Comput. Stat., 2010, 2, 433-459.

10. Rinnan A., van den Berg F., Engelsen S.B. Review of the most common pre-processing techniques for near-
infrared spectra, Trends Anal. Chem., 2009, 28:10, 1201-1222.

11. Geladi P. and Kowalski B.R., Partial least-squares regression: a tutorial, Anal. Chim. Acta, 1986, 185, 1-17.

Find out more at www.thermofisher.com/extrusion

©2017 Thermo Fisher Scientific Inc. All trademarks are the property of Thermo Fisher Scientific and its
subsidiaries unless otherwise specified. AN52307_E 12/17M
 APPLICATION NOTE AN52327

Measuring the equilibrium constant of a keto-enol

tautomerism using benchtop NMR
Author Abstract
Dean Antic, Thermo Fisher Scientific, The 1H NMR spectrum of a β-ketoester, ethyl acetoacetate, reveals the
Boulder, CO, USA presence of tautomerism, a dynamic process of intramolecular hydrogen
bonding. Proton exchange between the tautomeric species is slow, allowing for
Keywords direct measurement of each of the enol and keto forms of ethyl acetoacetate.
pharmaceutical, synthesis, kinetics, The signal areas for each species were then integrated. From which, the
tautomerism, equilibrium percent composition of each constitutional isomer was calculated and the
equilibrium constant was derived. In addition to qualitative structure elucidation,
Application benefits 1
H NMR also provides quantitative information to evaluate solution dynamics.
The ethyl acetoacetate 1H NMR
spectrum demonstrates the utility Introduction
of benchtop NMR in measuring an Ethyl acetoacetate (EAA), a β-ketoester, is a colorless liquid with a sweet,
intramolecular equilibrium. NMR is fruity aroma and flavor. Its primary application is as a starting material in the
ideal for studying reaction kinetics, as synthesis of α-substituted acetoacetic esters and cyclic compounds, as an
it is a nondestructive technique. An intermediate in the synthesis of vitamins and pharmaceuticals, and in the
advantage of the Thermo Scientific™ manufacture of inks, plastics, dyes and lacquers. Ethyl acetoacetate undergoes
picoSpin™ NMR spectrometer in self-enolization (Figure 1), where the keto form 1 exists in equilibrium with
monitoring reaction kinetics is that its enol isomer 2a and 2b. The process occurs because the α-hydrogen,
small quantities of reaction mixtures adjacent to the carbonyl, is slightly acidic, with pKa values ranging from 9-13 in
can be directly injected into the β-dicarbonyls, making ketones readily enolizable.1 This intramolecular chemical
capillary micro probe without the need equilibrium between keto and enol isomers is referred to as tautomerism. The
for dilution in a deuterated solvent. chemical equilibrium is thermodynamically driven. Typically, the more stable
keto form is favored at room temperature and above. However, in β-diketones,
stabilization of the enol form arises from stabilization by π-system conjugation.
Terminal groups and α-substituents also play a role in enol formation, where
bulky R groups in the α position results in steric effects destabilizing the enol,
while electron withdrawing groups in the α position help to stabilize it.

Figure 1: Keto-enol tautomerism of ethyl acetoacetate.

Tautomers can be readily studied by proton NMR methyl protons on C14 (e, C8; 3 JHH = 7.12 Hz) generates
because the high energy barrier for the interconversion a triplet due to spin-spin coupling to the adjacent
between enol-keto tautomers (Figure 1) makes it a slow methylene protons on C13 (e, C7). The methylene (CH2)
process that’s amenable to NMR analysis. Hence, keto resonance due to the protons on C13 appears as a spin-
and enol protons experience different local environments, coupled quartet centered at 4.14 ppm (3 JHH = 7.12 Hz).
resulting in easily identifiable proton chemical shifts.2 The enol tautomers 2a and 2b do not produce unique
For EAA, this is evident in the proton spectrum, where resonances separately because their conversion is a
resonances due to the keto form dominate.1-4 By rapid exchange process. At the opposite terminus, the
integrating these signals the equilibrium constant, Keq, acetyl methyl protons are affected by tautomerization due
and the %enol can be calculated for a given compound to changes in electron density at the carbonyl carbon
at a specific sample temperature. (e, C4), resulting in an upfield shift of the enol signal
(C9) relative to those of the keto signal (C18). The former
Experimental appears as a weak singlet at 1.94 ppm, while the latter
The 1H NMR spectrum of EAA was measured from is at 2.21 ppm. The α-protons of the keto and enol (C3)
a neat sample at 82 MHz and 32 °C using the forms manifest as singlets at 3.48 ppm (C15) and 5.03
Thermo Scientific™ picoSpin™ 80 spectrometer. The ppm (C3), respectively. The enol hydroxylic proton (C5a)
spectrometer’s capillary cartridge was fitted with micro- appears as an unsaturated alcohol, but intramolecular
fluidic inlet and outlet connectors that allow liquid sample hydrogen bonding causes it to behave like a carboxylic
injection into the spectrometer’s RF coil. The fluid path acid, shifting its resonance far downfield to 12.14 ppm.
was Teflon/Quartz capillary tubing with a total flow-
path volume of 40 microliters (μL). Liquid samples were The enol-keto equilibrium constant, Keq = [enol]/[keto],
introduced by manual injection using a disposable 1 mL and %enol are functions of temperature. Their values can
syringe and a 22-gauge blunt tip needle. be determined by integrating the enol and keto signal
areas. Table 1 summarizes the equilibrium constant and
Spectra were acquired using the following acquisition percent enol concentration values calculated for the
parameters: a 90° RF excitation pulse, a 1000 ms different combinations of enol to keto ratio. Note that
acquisition time, and a 12 second recycle delay. All each peak integration area is weighted according to
spectra were acquired with signal averaging. Spectral the number of protons that contribute to the respective
data were stored in a JCAMP-DX file format and imported signal. For example, the enol methyl contains three
into MestrelLab Research’s Mnova™ NMR analysis protons while the keto methylene has only two protons;
program for processing. Standardized data processing the weighting factor, in this case, is 2/3.
was applied across all spectra, specifically, by applying
zero filling, applying phase correction, and filtering using
exponential apodization. Resonance signals were fitted
using generalized Lorentzian functions, and integrated
using Mnova’s Global Spectrum Deconvolution method.

Results and discussion

Figure 2 shows the 1H NMR spectrum of ethyl
acetoacetate. Based on chemical shifts, multiplicity
and relative signal intensities, assignment of signals are
straightforward and unambiguous. The enol and keto
forms are indicated by ‘e’ and ‘k’, respectively, in Figure 2.
The signals were internally referenced relative to the large
singlet proton resonance at 2.21 ppm originating from the
acetyl methyl (C18) protons of the keto isomer.
Figure 2: Full 1H NMR spectrum of ethyl acetoacetate. Enol and keto
The chemical shift of ethyl ester protons (-OCH2CH3) are forms are indicated by ‘e’ and ‘k’, respectively. The prefix “et” refers to
unaffected by tautomerization and appear as classic ethyl. Inset shows the numbering system used to assign signals in the
NMR spectrum of ethyl acetoacetate.
spin-coupled signals. Centered at 1.23 ppm, the terminal
 Signal Keq %enol Signal Keq %enol Keq %enol
e
CH3 e
CH 0.09a 8.0%a,b
k
CH2 0.094 9.4% k
CH2 0.103 10.3%
k
CH3 0.096 9.6% k
CH3 0.105 10.5%
a
Ref. 4. b
Ref. 2.
Table 1. Equilibrium constant and percent tautomer composition measured at 32 °C.

From Figure 2, the calculated average Keq is 0.0992 References

±0.005 and %enol is 9.9%. Our values compare favorably 1. Allen, G.; Dwek, R. A. J. Chem. Soc. B, 1966,
to those reported by others,1,4 where the calculated 161-163.
values for Keq and %enol are 0.09 and 8.0%, respectively, 2. Cook, G.; Feltman, P. M. J. Chem. Educ. 2007, 84,
at 32 °C. These results confirm that the keto tautomer 1827-1829.
is the more favorable structure, and that the ethyl ester 3. Drexler, E. J.; Field, K. W. J. Chem. Educ. 1976, 53,
group doesn’t introduce additional steric interaction to 392-393.
destabilize the keto form. 4. Burdett, J. L.; Rogers, M. T. J. Am. Chem. Soc.
1964, 86, 2105-2109.
Conclusions
The 1H NMR spectrum of a β-ketoester, ethyl
acetoacetate, reveals the presence of tautomerism, a
dynamic process of intramolecular hydrogen bonding.
Proton exchange between the tautomeric species is
slow, allowing for direct measurement of each of the enol
and keto forms of ethyl acetoacetate. Integrating signal
areas for each species affords relative concentration
information to calculate the equilibrium constant and
percent composition of each constitutional isomer. In
addition to qualitative structure elucidation, 1H NMR
also provides quantitative information used to evaluate
solution dynamics.

Thermo Scientific picoSpin 80 NMR spectrometer

Find out more at www.thermofisher.com/nmr

©2017 Thermo Fisher Scientific Inc. All trademarks are the property of Thermo Fisher Scientific and its
subsidiaries unless otherwise specified. AN52327_E 12/17M
APPLICATION NOTE

No. AN52889
Benchtop NMR Combined with
GC/MS Confirms Identity of Forensic
Case Sample
Authors: Dean Antic, Ph.D.,
Thermo Fisher Scientific,
San Jose, CA, USA
WanLi Wei, Senior Engineer,
Institute of Forensic Sciences, Wuxi Public
Security Bureau, Wuxi, China

Key words
NMR, GC/MS, picoSpin, illicit drug, forensics The case sample demonstrated in this note exemplifies
the need for multiple techniques in order to confirm the
The benchtop Thermo Scientific™ picoSpin™ 80 1H identity of seized forensic samples, as recommend by
NMR spectrometer provides an additional layer of SWGDRUG.
structural identification of drug analogues and precursors,
complementing GC/MS analysis. The combination of Introduction
the two techniques allows for a positive identification of Illicit drugs identification is a challenge to law enforcement
real forensic case samples with high confidence, thereby due to the vast assortment of illegal drugs that already
enhancing the presumptive testing capabilities of illicit exist, and an increasing number of new, “not-yet” illegal
drug screening facilities. analogues of classified drugs appearing on the street.
Added to this burden is the identification of clandestine
Abstract lab chemicals and precursor compounds used in the
A forensic case sample of an illicit drug precursor was manufacture of illicit drugs, as well as the excipients used
analyzed by GC/MS and benchtop NMR. The most as adulterants to alter street-level drug purity.
probable chemical structure from a GC/MS library
search, however, conflicts with the spectral features Law enforcement has a variety of analytical tools at their
identified by NMR. By interrogating the peak pattern and disposal, including color tests, Fourier transform infrared
chemical shifts in the NMR spectra, both experimentally (FTIR), gas chromatography/mass spectrometry (GC/MS),
obtained and predicted by Mnova NMR software, the Raman spectroscopy and nuclear magnetic resonance
seized sample was identified as a compound with a spectroscopy (NMR) to aid in characterizing forensic
lower matching score by GC/MS. The combination of the samples and elucidating their structures.
two techniques allows one to discriminate between two
possible structural isomers in a real forensic case sample
with high confidence.
Since no single technique can provide definitive structure Seized sample identification methods
elucidation, the Scientific Working Group for the Analysis Figure 1 shows the GC/MS chromatogram of a sample
of Seized Drugs1 (SWGDRUG) proposes a combination of seized from a clandestine lab. The unknown sample was a
analytical techniques be used depending on the nature of clear, highly volatile, and fragrant organic liquid, suspected
the samples and available techniques for analysis. Initial as a precursor compound used in the production of illicit
chemical identification of a seized sample includes a drugs. The data was acquired using a Thermo Scientific™
color test and FTIR analysis. Molecular weight information ISQ™ QD Single Quadrupole GC-MS system. The main
is provided by mass spectrometry. When combined component of the sample has a retention time of 4.38 min
with other spectroscopic techniques, including Raman (top trace) and the corresponding mass spectrum is shown
spectroscopy and NMR, a chemical structure can be at the bottom of Figure 1. The main component has a
elucidated. The power of GC/MS, FTIR and Raman is molecular mass of 202 g/mol. In addition, there are a series
highly dependent upon the databases of reference samples of fragmentation peaks at m/z 159, 129, 103, 73, and 57.
and related compounds. It is, therefore, an arduous The subsequent library search yielded a high probability
endeavor to maintain and update these databases in the match of the sample to 1,1-dibutoxy butane, with its
face of increasing numbers of novel structural analogues structure shown in Figure 2A.
that are appearing.
Figure 1
ISQ QD GC-MS
chromatogram of unknown
liquid (forensic case sample).

Figure 2
Chemical structures of two
isomers; (A) 1,1-Dibutoxybutane;
A B and (B) 1,1-Diisobutoxy-2-
methylpropane.
The sample was further analyzed by a picoSpin 80 1H The spectrum exhibits a distinct pattern of multiplicities
NMR spectrometer, an 82 MHz pulsed, Fourier transform characteristic to the isopropyl group: two strong overlapping
1
H NMR permanent magnet instrument, equipped with a doublets centered near 0.95 ppm, followed by a more
capillary cartridge probe. Since the sample was a clear complex series of weaker multiplets between 1 – 2 ppm. In
organic liquid, dilution in typical NMR solvents was not addition, two strong doublets emerge centered at 3.34 and
required. A small amount of the case sample was placed in 4.49 ppm, respectively. The pattern of doublets suggests
a vial, to which a few drops of tetramethylsilane (TMS) was the presence of a single neighboring proton (CH), and the
added to reference chemical shifts. The mixture was then shift to high frequency indicates that the carbon center
directly injected into the NMR spectrometer. The resulting is attached to an electron withdrawing group. The NMR
spectrum is shown in Figure 3. Note that the peaks spectrum also implies structural symmetry of the sample.
attributed to impurities and TMS are manually labeled gray, These suggested structural features, however, conflict with
whereas the unknown compound peaks are labeled blue. 1,1-dibutoxybutane suggested by the GC/MS library search.

Figure 3
Experimental 1H NMR 82 MHz spectrum of the seized sample. Data was acquired using a 90° pulse and 15 s recycle delay between pulses. The spectrum is an average of 5 co-added
scans and processed using the Mnova NMR software suite (Mestrelab Research Inc.).
Figure 4 shows the comparison between the predicted The key features of the NMR spectrum are assigned as
NMR spectrum of 1,1-dibutoxybutane by Mnova software follows:
(top) and the experimental 1H NMR spectrum of the seized
• An intense, overlapping set of doublets at 0.9 ppm
sample (bottom). There are vast differences between the
is due to the terminal methyl’s of the isopropyl group
two spectra, indicating the sample under analysis is not
(-CH(CH3)2);
1,1-dibutoxybutane.
• A broad, weak multiplet between 1 - 2 ppm is due to the
Upon further examination of the compound list from the methine (CH) coupling to the adjacent CH3;
GC/MS library search, a low probability match compound,
• The doublet centered at 3.3 ppm originates from the
1,1-diisobutoxy-2-methylpropane (Figure 2B), was identified
methylene groups (CH2) of the 2-methyl propyl moiety.
as a strong candidate. The structure possesses all key
The doublet arises from the coupling to the adjacent CH
elements suggested by the experimental NMR spectrum:
proton of the isopropyl fragment;
isopropyl groups, electron withdrawing O atoms, and high
symmetry. Figure 5 shows a good agreement between the • The doublet at 4.5 ppm belongs to the CH group. It shifts
predicted NMR spectrum of 1,1-diisobutoxy-2-methylpropane further due to their attachment to two electron withdrawing
(top) and the experimental NMR spectrum of the seized O atoms. The weaker signal compared to the doublet at
sample (bottom). 3.3 ppm is due to the presence of only one CH group.

Figure 4
Predicted NMR spectrum of 1,1-dibutoxybutane (top) and
experimental 1H NMR 82 MHz spectrum of the seized
sample (bottom). The predicted spectrum was generated
by using Mnova NMRPredict plugin. Field strength was
set at 82 MHz and linewidth value was set at 2.5 Hz.

Figure 5
Experimental 1H NMR 82 MHz spectrum of the seized
sample and the predicted NMR spectrum of 1,1-diisobutoxy-
2-methylpropane generated by Mnova software.
The GC/MS data in combination with the picoSpin 80
NMR spectrum allows for the positive identification of
the seized sample as 1,1-diisobutoxy-2-methylpropane.
The compound is a precursor chemical and not a variant
structural analogue of known illegal drugs.
Thermo Scientific™ picoSpin™ 80
NMR spectrometer and Thermo
Conclusions Scientific™ ISQ™ QD Single
In this note we showed how the picoSpin 80 NMR was Quadrupole GC/MS System
able to provide additional information by interpreting
and predicting the likely structure of a seized sample,
complementing those derived from GC/MS analysis. The
combination of the two techniques allows for a positive
identification of an unknown compound with high confidence.
The practices demonstrated in this note conform to the
recommendations by SWGDRUG that multiple techniques
are required to confirm seized sample identity.

The benchtop picoSpin 80 1H NMR solution offers

structure selectivity and discriminating power needed to
provide an additional layer of structural identification of
drug analogues and precursors. This compact instrument
can be conveniently placed in workspace-limited labs and
testing areas, while enhancing the presumptive testing
capabilities of illicit drug screening facilities.

References
SWGDRUG home page: http://www.swgdrug.org/
(accessed Jun 17, 2016).

In the United States: In Canada:

For customer service, call 1-800-766-7000 For customer service, call 1-800-234-7437
To fax an order, use 1-800-926-1166 To fax an order, use 1-800-463-2996
To order online: thermofisher.com To order online: thermofisher.ca

Find out more at www.thermofisher.com/nmr

For Research Use Only. Not for use in diagnostic procedures. © 2016 Thermo Fisher Scientific Inc. All rights
reserved. Mnova is a trademark of Mestralab Inc. All other trademarks are the property of Thermo Fisher Scientific
and its subsidiaries unless otherwise specified. AN52889_E 08/16M
APPLICATION NOTE

No. AN52907
Determination of polymer molecular weight and
composition using picoSpin NMR spectroscopy
Authors:
Katherine Paulsen and Daniel Frasco,
Thermo Fisher Scientific, Madison, WI

Key words

NMR, polymer, copolymer, number average molecular

weight, compositional analysis

Abstract
Polymer molecular weight determination and copolymer by comparing the resonance signals from repeating
compositional analysis involve the integration of the units to those of end groups, thereby determining the
resonance signals from polymer repeating units and end molecular weight. A similar methodology can be adapted
groups, which are inherently broad. A Thermo Scientific™ for the compositional analysis of copolymers, so long as
picoSpin™ 80 NMR spectrometer is well suited for these the resonance signals from different monomers can be
analyses. In the example of poly (ethylene glycol) (PEG) clearly differentiated.
acetyl triarm, the number average molecular weight
was determined with great ease. In the case of the Due to poor molecular rotation and marginally different
compositional analysis of a commercial PEG-PPG-PEG chemical environments in which the repeating units are
block copolymer, Pluronic® L-35, the PEG/PPG ratio situated, resonance signals from polymer repeating units
determined by a picoSpin 80 NMR agrees well with the often coalesce as a broad peak, even using high-field
manufacturer’s product specification. Furthermore, the NMR spectrometers.3 However, in the molecular weight
analyses using an 82 MHz and a 300 MHz NMR yielded analysis and compositional analysis, collective signals
almost identical results. from polymer repeating units are used for calculation
and monomeric resolution is generally not required. To
Introduction that end, low-field NMR readily lends itself as a low-cost
The control of the molecular weight and molecular weight alternative to high-field instruments, with significant savings
distribution (MWD) is essential to obtain and improve on both instrument procurement and upkeep.
certain desired physical properties in a polymer product,
and is therefore of great importance in material science1. In this application note, molecular weight determination of
In addition to gel permeation chromatography (GPC) poly(ethylene glycol) (PEG) acetyl triarm and compositional
and matrix assisted laser desorption ionization mass analysis of a polyol using a picoSpin NMR spectrometer
spectrometry (MALDI MS)2, end group analysis using are presented. The results of the compositional analysis
nuclear magnetic resonance (NMR) spectroscopy has using an 82 MHz NMR were also compared with those
long been established as a viable analytical technique using a high-field, 300 MHz NMR spectrometer.
for determining the number average molecular weight
(Mn) of polymers. For polymers with a defined end group
structure, the number of repeating units can be deduced
Experimental Results and discussion
The structures of the polymers examined in this study are Number average molecular weight (Mn) determination.
shown in Figure 1. In 1H NMR spectroscopy, the area under each resonance
signal is proportional to the molar concentration of the
A B protons being analyzed. Number average molecular
weight (Mn) determination by end-group analysis using
1
H NMR therefore involves identifying and integrating
distinguishable proton signals originating from end-
groups and repeating units.

Figure 1: Chemical Structures of: A) poly(ethylene glycol) (PEG) acetyl triarm (PEG acetyl
triarm), and B) poly(ethylene glycol)-block-poly(propylene glycol)-block-poly(ethylene
glycol) (PEG-PPG-PEG).

A Thermo Scientific picoSpin 80 NMR spectrometer was

used to acquire the spectra of the polymer materials. The
spectrometer is an 82 MHz, pulsed, Fourier transform
1
H NMR. The instrument contains a 2 Tesla temperature
controlled permanent magnet and is fitted with a 40
microliter capillary cartridge used for sample introduction
into the spectrometer.

Poly(ethylene glycol) (PEG) acetyl triarm (PEG acetyl triarm)

was a crude reaction mixture provided by Dr. Lei Zhu’s
group in the Department of Macromolecular Science and Figure 2: 82 MHz 1H NMR spectrum of PEG acetyl triarm reaction mixture.
Engineering at Case Western Reserve University.
Figure 2 shows the 1H NMR spectrum of PEG acetyl
The sample was injected neat and 64 scans were acquired. triarm obtained using an 82 MHz picoSpin NMR. The
Poly(ethylene glycol)-block-poly(propylene glycol)-block- acetyl (-OCOCH3) end-group protons resonate as a sharp
poly(ethylene glycol) (PEG-PPG-PEG; Pluronic® L-35) was singlet at δ 1.8 ppm with a normalized peak area of 1.00.
purchased from Sigma-Aldrich.4 The sample was diluted The three acetyl moieties contain a total of 9 protons.
in equal volume of CHCl3 prior to injection. The data The PEG repeating group (-OCH2CH2-) protons and the
was collected using 16 scans. The 300 MHz NMR data glycerol protons resonate between δ 3.1-3.8 ppm with
of Pluronic L-35 diluted in CDCl3 was acquired at Case a normalized peak area of 9.65. Each repeating group
Western Reserve University. of PEG contains 4 protons, and the glycerol linkage
(–CH2CHCH2–) contains 5 protons. Since there are
The liquid samples were injected into the capillary cartridge three arms for the polymer, the total number of protons
using 1 mL slip-tip polypropylene syringes and 22 gauge contributing to the signal between δ 3.1-3.8 is 12n+5.
blunt-tipped needles. Back-to-back sample injection and
data acquisition was separated by a solvent/air/solvent/air peak area of repeating units and glycerol linkage peak area of end groups
=
# of protons in repeating units and glycerol linkage # of protons in end groups
flush to thoroughly clean the capillary. Samples were either
5.78 =.>>
referenced to chloroform (CHCl3, δ7.26 ppm) or another = , n = 6.82 ≈ 7 repeating monomer units
12n + 5 5

known signal in the spectrum. MF =FWIJKLMNOJ JPFQRIM +FWMFS INOTUV + FWrepeating units =41+177+(44×7×3)=1142

integral of methyl protons 3.00
All spectra were acquired using the following acquisition Based on the end
x = group analysis using = an 82 MHz
# of methyl group protons 3
parameters: 90° excitation pulse, 750 ms acquisition time picoSpin NMR x = 1.0 relative moles of propylene glycol
spectrometer, the number average
and 5 second recycle delay. Spectral data were processed molecular weight of the PEG acetyl triarm is 1142 g/mol.
using the Mnova™ NMR analysis software with a standard 7.65 3.00
=
set of processing parameters including: Zero filling and 4y + 3 3
y = 1.16 relative moles of ethylene glycol
phase correction. Apodization was not used.


relative moles PEG×MWefg
weight % PEG = ×100%
relative moles PEG×MWefg + relative moles PPG×MWeeg
1.16×44
= ×100% = 46.8%
1.16×44 + 1.0×58
Compositional analysis of a PEG/PPG block copolymer.
Figure 3 shows the 1H NMR spectra of Pluronic L-35, with a normalized peak area of 3.0, and the CH- and
a PEG-PPG-PEG copolymer, using (a) an 82 MHz CH2-proton signals are located at δ ~ 3.3-3.7 ppm with
picoSpin NMR spectrometer; and (b) a 300 MHz NMR a normalized peak area of 7.77. The minor variances in
spectrometer. The two spectra are very similar. In the chemical shift between the two spectra are likely due
82 MHz spectrum, the signals at δ ~0.8 ppm with a to the difference in sample concentrations. Because of
normalized peak area of 3.0 are attributed to the methyl the higher sensitivity of a 300 MHz NMR instrument,
(-CH3) group of PPG. The CH- and CH2- proton signals the samples were used in a lower concentration; hence,
from both PPG and PEG are located at δ ~ 3.1-3.4 ppm, the difference in chemical shift. The 300 MHz spectrum
with a normalized peak area of 7.65. The contributions also offers greater resolution than the 82 MHz one. The
to these signals include 3 protons from PPG (one CH- increased resolution, however, should have little to no
and one CH2-) and 4 protons from PEG (2 CH2- groups). bearing on the results of the compositional analysis,
The 300 MHz 1H NMR spectrum is essentially the same: because the collective signals are integrated for the
the signal from the methyl group resonates at δ 1.1 ppm ensuing calculations.

A Figure 3: A) 82 MHz 1H NMR spectrum of PEG-PPG-PEG

prepared 50:50 (v/v) in CHCl3. B) 300 MHz 1H NMR spectrum
of PEG-PPG-PEG prepared in CDCl3.

B
 A detailed compositional analysis using the 82 MHz NMR Conclusions
spectrum is outlined below.
peak area of repeating units and glycerol linkage
=
peak area of end groups

1
H NMR spectroscopy has been established as a powerful
# of protons in repeating units and glycerol linkage # of protons in end groups
tool for polymer characterization, including molecular
5.78 =.>>
= , n = 6.82 ≈ 7 repeating monomer units
1. peak area of repeating units and glycerol linkage
Determine the
12n + 5 5relative moles of PPG, denoted as x,
peak area of end groups
weight determination and copolymer compositional
MF =FWIJKLMNOJ JPFQRIM +FWMFS INOTUV + FWrepeating units=
=41+177+(44×7×3)=1142
using the signals at δ ~0.8.
# of protons in repeating units and glycerol linkage
peak area of repeating units and glycerol linkage # of protons in end groups
peak area of end groups
analysis. Both analyses involve the integration of the
=
5.78 =.>>
# of protons in repeating units and glycerol linkage
= # of protons in end groups
, n = 6.82 ≈ 7 repeating monomer units resonance signal from polymer repeating units, which are
12n + 5 5
integral of methyl protons 3.00
5.78
MF =FWIJKLMNOJ JPFQRIM
12n + 5 +FW
=.>>
=x = MFS INOTUV
, n = 6.82 =
≈ 7 repeating monomer units
+ FW repeating units =41+177+(44×7×3)=1142
5# of methyl group protons 3
inherently broad. The Thermo Fisher Scientific picoSpin 80
MF =FWIJKLMNOJ JPFQRIMx = 1.0 relative moles of propylene glycol
+FWMFS INOTUV + FWrepeating units =41+177+(44×7×3)=1142 NMR spectrometer is well suited for these investigations.
integral of methyl protons 3.00 In the example of poly(ethylene glycol) (PEG) acetyl triarm,
x = =
2. Determine the # of methyl group protons
relative moles3.00 of PEG,
integral of methyl protons
7.65 3denoted as y,
3.00 the number average molecular weight was determined with
x = = =
x = 1.0 relative moles of propylene glycol
# of methyl group protons 3
using the signals at δ4y + 3~3.1-3,4. 3
great ease. In the case of the compositional analysis of a
x = 1.0 relative moles of propylene glycol
y = 1.16 relative moles of ethylene glycol
commercial PEG-PPG-PEG block copolymer, Pluronic®
7.65 3.00
4y + 3
7.65
=
3
3.00
L-35, the PEG/PPG ratio determined by a picoSpin

=
y = 1.16 relative moles of ethylene glycol
4y + 3 3
relative moles PEG×MW
80 NMR agrees well with the manufacturer’s product
efg
weight % PEG = ×100%
y = 1.16 relative moles of ethylene glycol
relative moles PEG×MW efg + relative moles PPG×MWeeg specification. Furthermore, the analyses using an 82 MHz
3. Calculate
=
the weight
1.16×44 percentages of PEG and PPG.
×100% = 46.8%
and a 300 MHz NMR yielded almost identical results. The
1.16×44 + 1.0×58 relative moles PEG×MWefg

weight % PEG = ×100% Thermo Fisher Scientific picoSpin 80 NMR spectrometer
relative moles PEG×MW efg + relative moles PPG×MWeeg
relative moles PEG×MW
weight % PEG =
1.16×44 relative moles PPG×MW
efg
×100% has proven to be a low-cost alternative to high-field NMR
relative moles PEG×MW efg + relative moles PPG×MW
eeg eeg ×100%
weight % PPG =
= ×100% = 46.8%
1.16×44 relative moles PEG×MW
+ 1.0×58
1.16×44 efg + relative moles PPG×MWeeg spectrometers for polymer molecular weight determination
= ×100% = 46.8%

=
1.0×58
1.16×44 + 1.0×58
×100% = 53.2% and compositional analysis.
1.16×44 + 1.0×58 relative moles PPG×MWeeg
weight % PPG = ×100%
relative moles PEG×MW efg + relative moles PPG×MWeeg
relative moles PPG×MW
weight % PPG =
1.0×58
eeg
relative moles PEG×MWefg + relative moles PPG×MWeeg
×100% References
= ×100% = 53.2%
1.16×44 + 1.0×58
1.0×58
= ×100% = 53.2%
1.16×44 + 1.0×58 1. Higginbotham, C. L.; Izunobi, J.U. J. Chem. Ed. 2011,
88, 1098-1104
Based on the 82 MHz NMR spectrum, Pluronic L-35
copolymer contains 46.8 % of PEG and 53.2 % of PPG. 2. Skoog, D. A.; Holler, J. F.; Crouch, S. R. Principles
The results using the 300 MHz spectrum are essentially the of Instrumental Analysis, 6th ed.; Brooks Cole: Pacific
same: 47.4 % of ethylene glycol and 52.6 % of propylene Grove, 2006.
glycol. Both results are in good agreement with the product
specification4, where PEG/PPG (w/w) = 1. 3. Bovey, F. A.; Mirua, P. A. NMR of Polymers, 1st ed.;
Academic Press: San Diego, 1996.

4. Sigma-Aldrich. Poly(ethylene glycol)-block-

poly(propylene glycol)-block-poly(ethylene glycol) (PEG-
PPG-PEG; Pluronic L-35; CAS No. 9003-11-6).
www.sigmaaldrich.com/catalog/product/aldrich/435414

Thermo Scientific picoSpin 80 NMR spectrometer

Find out more at www.thermofisher.com/nmr

For Research Use Only. Not for use in diagnostic procedures. © 2016 Thermo Fisher Scientific Inc. All rights
reserved. Mnova is a trademark of Mestralab Inc. All other trademarks are the property of Thermo Fisher Scientific
and its subsidiaries unless otherwise specified. AN52907_E_11/16M
APPLICATION NOTE

No. AN52948
Qualitative and quantitative analysis of
the polymerization of PS-b-PtBA block
copolymer using picoSpin 80 NMR
Authors:
Katherine Paulsen and Daniel Frasco,
Thermo Fisher Scientific, Madison, WI, USA
Yufeng Zhu, Department of
Macromolecular Science and Engineering,
Case Western Reserve University,
Cleveland, OH, USA
Key words
picoSpin 80, NMR, polymerization, copolymer, number of
repeating units, polystyrene

Benefit
The Thermo Scientific™ picoSpin™ 80 NMR spectrometer
offers near real-time analytical capability for reaction
monitoring due to its unique capillary injection design Introduction
and compact footprint. The picoSpin 80 NMR has also Nuclear magnetic resonance (NMR) spectroscopy has
proven to be a low-cost alternative to high-field NMR proven to be an invaluable analytical tool for reaction
spectrometers to quantitatively assess polymerization and monitoring and optimization by elucidating molecular
to determine copolymer structure. structure, studying reaction kinetics, monitoring reaction
progress and gauging product purity1. For polymers
Abstract however, resonance signals from repeating units often
The use of the low-field picoSpin 80 NMR spectrometer coalesce as broad peaks, even with high-field NMR
offers a near real-time analytical capability for reaction spectrometers. This is largely due to poor molecular
monitoring, largely due to its unique capillary injection design rotation and repeating units being situated in marginally
and compact footprint. Despite the broad resonance signals different chemical environments. Low-field NMR such as
typical for polymers, by carefully selecting the regions where the picoSpin 80 NMR spectrometer lends itself as a low-
signal changes manifest the involved chemical transformation, cost alternative to high-field instruments for polymerization
the integrated peak area provides valuable insight to the monitoring, resulting in significant cost savings on both
reagent conversion as well as the final product structure. instrument procurement and upkeep.
Presented herein is a case study using a picoSpin 80 set of processing parameters including zero filling and
NMR spectrometer to monitor the polymerization of t-butyl phase correction. Apodiation was not used.
acrylate with a polystyrene reagent. The reaction scheme
is shown in Figure 1. By carefully identifying and integrating Results and discussion
the resonance signals associated with the monomer, the A qualitative look at the polymerization
progress of the polymerization was successfully assessed Figure 1 depicts a narrowly dispersed polystyrene-based
both qualitatively and quantitatively. The number of t-butyl reversible addition-fragmentation chain transfer (RAFT)
acrylate units incorporated into the final copolymer product reagent reacting with t-butyl acrylate monomer to form a
was also deduced. PS-b-PtBA block copolymer2-3. Figure 2 shows the NMR
spectrum of the initial reaction mixture containing t-butyl
acrylate (tBA) and polystyrene reagent (PS) in anisole. The
signals between δ 1.4–2.8 ppm originate from the 3 protons
of the CH2- and CH- groups of the PS repeating unit. These
overlap with the singlet from the 9 t-butyl methyl protons
Figure 1: Polymerization of t-butyl acrylate with a polystyrene reagent to form PS-b-PtBA. of tBA (δ 1.7 ppm). The singlet at δ 3.7 ppm is ascribed to
the methyl group of anisole. The signals between δ 5.5 –
Experimental 6.7 ppm correspond to the 3 vinyl protons of tBA. Finally,
Monomer t-butyl acrylate (tBA, 0.150 g, 1.170 mmol), the multiplet in the region δ 6.8 – 7.6 ppm arises from the
polystyrene reagent (PS, 0.230 g, 0.120 mmol), 2,2’-azobis aromatic rings from both anisole and PS.
(2-methylpropionitrile) radical initiator (AIBN, 0.001 g, 0.006
mmol), and anisole (1.156 mL) were added together into a
10 mL flask. The number average molecular weight of PS
is 1910.76 g/mol, determined by the end-group analysis
using a 500 MHz NMR and confirmed by GPC. The
average number of repeating units in PS is 16. This resulted
in a molar ratio of 9.75/1 for tBA/PS. The mixture was
degassed by three freeze-pump-thaw cycles. The mixture
was then stirred at 75 °C for 12 h and the polymerization
process was observed by 1H NMR spectroscopy.

A picoSpin 80 NMR spectrometer was used to acquire

spectra of the polymerization reaction. The instrument is an
82 MHz, pulsed, Fourier transform 1H NMR spectrometer Figure 2: 82 MHz 1H NMR spectrum of the initial reaction mixture containing polystyrene
with superior chemical shift resolution and sensitivity. reagent (PS) and t-butyl acrylate (tBA) in anisole.

The instrument contains a 2 Tesla temperature controlled

permanent magnet and is fitted with a 40 microliter capillary
cartridge used for sample injection into the spectrometer.

All samples were manually injected into the capillary

cartridge using disposable 1 mL slip-tip polypropylene
syringes and 22 gauge blunt-tipped needles. During the
reaction, aliquots were taken out and injected directly into
the picoSpin capillary cartridge with no sample preparation.
Back-to-back sample injection and data acquisition was
separated by a solvent/air/solvent/air flush of the capillary
cartridge with anisole used as the solvent. Spectra were
referenced to the –CH3 signal from anisole (δ3.75 ppm).

All spectra were acquired using the following acquisition

parameters: 90° excitation pulse, 750 ms acquisition time
and 8 s recycle delay. The spectral data was processed Figure 3: Overlay of the 82 MHz 1H NMR spectral collected at 0, 1, 2, 3, and 12 hour
using the Mnova™ NMR analysis program with a standard time-points. The insert shows a decrease in intensity of the vinyl protons of tBA.
Figure 3 shows the overlaid 1H NMR spectra of the reaction
mixture acquired at 5 different time points over the course
of the polymerization: 0, 1, 2, 3 and 12 hours, respectively.
The main chemical transformation during the polymerization
is the opening of the double bonds in the tBA monomer,
as demonstrated in Figure 4. The insert in Figure 3 clearly
demonstrates that the resonance signals from the vinyl
protons of tBA (δ 5.5–6.7 ppm) decrease as the reaction
progresses. A quantitative perspective of the chemical
transformation is presented in the subsequent section.

Figure 5: 82 MHz 1H NMR spectrum of the polymerization process after 12 hours at 75 °C.

For the vinyl protons in tBA (a total of 3 × 9.75 = 29.25),

assuming n% was converted to polymer, 29.25n protons is
shifted to the region δ 1.4–2.8 ppm and 29.25 (1-n) protons
remain in the region δ 5.5-6.7 ppm.

Figure 4: Schematic showing the conversion of vinyl protons to protons in the polymer
backbone. The numbers denote the projected chemical shifts of the involved protons by
Based on the calculations below, it was determined that
1
H NMR. 34% of tBA was converted into the block copolymer,
corresponding to an average of 34% of 9.75 = 3 tBA units
It is worth pointing out that the compact footprint of the in every copolymer molecule.
picoSpin 80 NMR spectrometer allows the instrument
to be set up in the laboratory where the chemistry takes
place. Furthermore, the unique capillary injection system
of the picoSpin minimizes sample workup, offering a
near real-time analytical capability critical for reaction
monitoring and optimization.

Quantitative Analysis of the Polymerization

Figure 5 shows the 1H NMR spectra of the reaction mixture
at t=12 hours. As the reaction progressed, the monomer
tBA was gradually added to the polymer through the Conclusion
opening of the double bonds (Figure 4). Consequently, The use of the low-field picoSpin 80 NMR spectrometer
part of the protons originally contributing to the resonance offers a near real-time analytical capability for reaction
signals in the region δ 5.5-6.7 ppm were converted to the monitoring, largely due to its compact footprint and unique
protons contributing to signals between δ 1.4 – 2.8 ppm. A capillary injection design. Despite the broad resonance
I
closer analysis reveals that the ratio 1.4–2.8 ppm increased signals typical for polymers, by carefully selecting the regions
I
from 4.88 at t=0 to 7.57 at t=12 hour. 5.5–6.7 ppm where signals can quantitatively manifest the involved
chemical transformation, the integrated peak area provides
At t=0, there are 9.75 moles of tBA for every mole of PS. valuable insight to the reagent conversion as well as the
At t=12, the signals between δ 1.4–2.8 ppm include the final product structure. In the case of the polymerization of
aliphatic protons from PS, the methyl protons from the PS-b-PtBA, the vinyl and aliphatic regions were selected for
t-butyl groups in tBA, and the protons converted from the integration. By comparing the signal ratio before and after the
vinyl groups in tBA. There are 3 aliphatic protons in each reaction, it was determined that 34% of the tBA monomer
PS repeating unit, and there are 16 repeating units in every was converted to the copolymer. On average, the final
PS molecule. The total PS protons are therefore 48. The product contains 3 tBA units. The described workflow and
number of methyl protons in tBA are 9 × 9.75 = 87.75 and its underlying principles should be of interest for chemists
remains unchanged throughout the reaction. routinely performing polymer synthesis.
References
1. Higginbotham, C. L.; Izunobi, J.U. J. Chem. Ed. 2011,
88, 1098-1104

2. Li, B.; Shi, Y.; Zhu, W.; Fu, Z.; Yang, W. Polym. J. 2006,
38, 387-394

3. Chiefari, J.; Chong, Y. K.; Erocle, F.; Krstina, J.; Jeffery,

J.; Le, T. P. T.; Mayadunne, R.T.A.; Meijs, G. F.; Moad, C.
L.; Moad, G.; Rizzardo, E.; Thang, S. H. Macromolecules
1998, 31, 5559-5562.

Find out more at thermofisher.com/picospin

For Research Use Only. Not for use in diagnostic procedures. © 2017 Thermo Fisher Scientific Inc. All rights
reserved. All trademarks are the property of Thermo Fisher Scientific and its subsidiaries unless otherwise specified.
AN52948_E 03/17M
APPLICATION NOTE

No. AN52976
Analysis of acetanilide herbicides and their
rotational isomers by picoSpin 80 NMR
Author:
Daniel Frasco, Thermo Fisher Scientific,
Madison, WI, USA
Key words: NMR, agrochemical,
pesticide, herbicide, isomers

Application benefit
Thermo Scientific™ picoSpin™ 80 Series II NMR is a
powerful analytical tool to qualitatively and quantitatively Due to the widespread use of these herbicides, it is
analyze acetanilides and their isomers. Compared to the important to understand the metabolic pathways of these
traditional methods using chromatography and high-field molecules and identify their transformation products. Special
NMR, the NMR method using picoSpin 80 is cost-effective, attention has to be paid to stereoisomers, as they could
time-efficient and straightforward. have vastly different reactivity and degradation pathways.3-4
For acetanilides, amide cis-trans isomerization has been well
Abstract established where the rotation of the amide bond from the
The picoSpin 80 Series II NMR spectrometer can be used as a substituted aromatic ring is hindered.3-6
low-cost alternative to a high-field instrument for the qualitative
and quantitative analysis of acetanilide herbicides. The Quantitative determination of the isomeric ratio in a mixture
picoSpin offers adequate resolution to differentiate between presents a particular analytical challenge. Traditionally, a
acetanilide isomers and a facile means to determine the number of techniques including Gas Chromatography-Mass
isomeric ratio by simply integrating the corresponding signals. Spectrometry (GC-MS), Capillary Zone Electrophoresis (CZE),
High-Performance Liquid Chromatography (HPLC) and
Introduction high-field Nuclear Magnetic Resonance (NMR), have been
Acetanilides are a class of used to analyze the acetanilide isomers.4-5 Chromatographic
preemergent herbicides that inhibit techniques including GC-MS, HPLC and CZE often require
the growth of unwanted grasses time-consuming method development. The chromatography
and weeds associated with the runs are generally on the order of 10-15 min, often with
production of agricultural crops like Figure 1: General Structure of non-baseline resolution, giving rise to the uncertainty in
corn, soybeans, cotton and peanuts.1 Acetanilide Herbicides isomeric ratio. High-field NMR has proven to be a powerful
The general structure of acetanilide tool for the analysis of acetanilide isomers by integrating their
herbicides is shown in Figure 1. signals. However, the associated intensive capital and space
investment often prohibits their use in smaller facilities or in
Acetanilides represent some of the most widely used factories where pesticide production is taking place.
pesticides for both crop and non-crop use. In 2007,
metolachlor and acetochlor, two representative acetanilides,
were the 4th and 5th most used pesticides in the United
States, with an approximated use of around 30 million
pounds each.2
This work demonstrates the qualitative and quantitative
analyses of a group of acetanilides with varying structural
complexity using an inexpensive and portable picoSpin 80
Series II NMR spectrometer. NMR spectra were obtained
in as little as a few seconds compared to the longer
analysis times of other techniques. Additionally, a facile
means to determine the isomeric ratio is also presented.
This methodology was applied to investigate the impact of
solvent polarity on the ratio of cis-trans rotational isomers
of the acetanilide.

Experimental
A picoSpin 80 Series II NMR spectrometer was used
to acquire the spectra of acetanilide herbicides. The Figure 2: 1H NMR spectrum of a 1.8M solution of propachlor in DMSO-d6 acquired in 5
spectrometer is an 82 MHz, pulsed, Fourier transform minutes using 32 scans. The residual DMSO signal and water present in the deuterated
solvent are identified.
1
H NMR. The instrument contains a 2 Tesla temperature
controlled permanent magnet and is fitted with a 40 in Figure 3 clearly indicates that the resolution of the picoSpin
microliter capillary cartridge used for sample introduction 80 Series II is sufficient to resolve the signals corresponding
into the spectrometer. to each isomer. In the case of alachlor, ethyl groups are
located at the 2 and 6 position of the aromatic ring. The
All samples were purchased from Sigma Aldrich and were orientation of the carbonyl group away from the aromatic ring
used as received. Both solid and liquid samples were corresponds to the trans- isomer, and the one aligning toward
analyzed in deuterated solvent and injected into the capillary the aromatic ring corresponds to the cis- isomer. In Figure 3,
cartridge using 1 mL slip-tip polypropylene syringes and 22 the three methoxy protons from the two isomers can be seen
gauge blunt-tipped needles. Back-to-back sample injection as two singlets at δ 3.3 and 3.4 ppm. The two methylene
and data acquisition was separated by a solvent/air/solvent/ protons next to the carbonyl show a significant difference in
air flush to thoroughly clean the capillary. chemical shift due to their change in proximity to the aromatic
ring in the two isomers. They are present as singlets at δ 3.8
All spectra were acquired using the following acquisition ppm (trans-) and δ 4.7 ppm (cis-). The two methylene protons
parameters: Between 1 and 64 scans, 90° excitation pulse, located between the amide nitrogen and the methoxy oxygen
750 ms acquisition time and 8 second recycle delay. Spectral resonate at δ 4.9 ppm for both isomers, although a small
data was processed using the Mnova™ NMR analysis shoulder ascribed to the minor cis- isomer can be observed,
software with a standard set of processing parameters. indicating a subtle difference caused by the isomerism.

Results and discussion

Figure 2 depicts the 1H NMR spectrum of propachlor
as a 1.8M solution in deuterated DMSO. The signals
corresponding to each proton can be readily assigned.
The doublet resonating at δ 1.0 ppm corresponds to the
two methyl groups of six protons, the singlet at δ 3.8 ppm
originates from the two methylene protons alpha to the
carbonyl, the septet at δ 4.8 ppm is from the one isopropyl
methine proton, and the five protons from the aromatic ring
are located between δ 7.1-7.6 ppm.

Substitutions adjacent to the amide bond on the aromatic

ring could hinder the rotation of the amide bond, resulting
in rotational isomers (cis-trans isomers). This phenomenon
Figure 3: 1H NMR spectrum of a 3.7M solution of alachlor in DMSO-d6 acquired using 32
has been extensively studied using high-field NMR scans. Signals labeled as 1 correspond to the trans isomer and 2 correspond to the cis
spectrometers; although the 1H NMR spectrum of alachlor isomer. Unlabeled multiplets represent an overlap of trans and cis signals.
The protons corresponding to the ethyl functional groups
resonating between δ 1.0-1.3 ppm and δ 2.4-2.7 ppm and
the aromatic signals located at δ 7.1-7.5 ppm cannot be
differentiated between the two isomers.

Figure 4 shows the NMR spectra of three acetanilides:

butachlor, alachlor and acetochlor. All three structures have
substitution at the 2- and 6- position of the aromatic ring,
leading to cis-/trans isomerism. Compared to alachlor,
both butachlor and acetochlor have longer alkyl side
chains attached to the oxygen. In addition, acetochlor has
unsymmetrical ring substituents. These structural differences
in butachlor and acetochlor compared to alachlor are
annotated in their respective 1H NMR spectra (Figure 4.) Figure 4: Stacked 1H NMR spectra of butachlor, alachlor and acetochlor in DMSO-d6. The
structural changes of butachlor and acetochlor compared to alachlor are highlighted.

Figure 5 shows the NMR spectra of alachlor in three different

solvents. As the solvent polarity decreases from DMSO-d6
to CDCl3, the isomeric ratio (trans/cis) increases. The results
are also summarized in Table 1. Using the most nonpolar
solvent CDCl3, the trans to cis ratio was 92.2% to 7.8%,
which was similar to previously reported data in CDCl3.4

The dependence of alachlor isomer distribution on solvent

polarity has been studied previously using high-field NMR
instruments, and similar trends have been reported.3-4 For
alachlor and other similar molecules with amide cis-trans
isomerization, the transition state between the two
isomers is characterized by a partial C-N bond rotation.
This transition state for amide bond rotation is less polar
compared to its cis and trans isomers. In more non-polar Figure 5: 1H NMR spectra of a 1M solution of alachlor using three different deuterated solvents.
solvents, the energy of the transition state decreases; thus,
promoting the isomerization. Since the trans isomer is more Relative % Trans % Cis
Solvent
stable than the cis isomer,3, 6 the trans/cis ratio increases as Polarity Isomer Isomer

solvent polarity decreases. DMSO-d6 0.444 73.7 26.3

Acetone-d6 0.355 83.6 16.4

This data confirms that the choice of solvent is critical to
CDCI6 0.259 92.2 7.8
the isomer distribution of acetanilide herbicides in solution,
and that the picoSpin 80 Series II NMR spectrometer can
Table 1: The effect of solvent polarity on isomer distribution in alachlor. All samples were
easily examine this ratio quantitatively. analyzed in triplicate and the average values are reported.

Conclusions
The picoSpin 80 Series II NMR spectrometer can be used
as a low-cost alternative to a high-field instrument for the
qualitative and quantitative analysis of acetanilide herbicides.
Some acetanilide molecules are present as rotational
isomers due to hindered rotation of the amide bond.
The picoSpin offers adequate resolution to differentiate
the isomers and a facile means to determine the isomeric
ratio by simply integrating the corresponding signals. The
results clearly suggest that solvent polarity has a profound
impact on the isomeric distribution in solution, and great
caution should therefore be exercised when analyzing and
interpreting the isomeric ratio.

References
1. Morton, M. D.; Walters, F. H.; Aga, D. S.; Thurman, E. M.;
picoSpin 80 Series II NMR spectrometer
Larive, C. K. J. Agric. Food Chem. 1997, 45, 1240-1243

2. Pesticides Industry Sales and Usage: 2006 and

2007 Market Estimates. United States Environmental
Protection Agency, 2011, https://www.epa.gov/
sites/production/files/2015-10/documents/market_
estimates2007.pdf. Accessed 3 January 2017.

3. Schmidt, W. F.; Hapeman, C. J.; Waters, R. M.; Johnson,

W. E. J. Agric. Food Chem. 1995, 43, 1724-1729

4. Aga, D. S.; Heberle, S.; Rentsch, D.; Hany, R.; Muller, S.

R. Environ. Sci. Technol. 1999, 33, 3462-3468

5. Cardoza, L. A.; Cutak, B. J.; Ketter, J.; Larive, C.K.

J. Chromatogr. A 2004, 1022, 131-137

6. Harrison, R. K.; Stein, R. L. J. Am. Chem. Soc. 1992,

114, 3464-3471

Find out more at thermofisher.com/nmr

For Research Use Only. Not for use in diagnostic procedures. © 2017 Thermo Fisher Scientific Inc. All rights
reserved. All trademarks are the property of Thermo Fisher Scientific and its subsidiaries unless otherwise specified.
AN52976 0517
APPLICATION NOTE AN52979

Hydrogen content determination by picoSpin NMR

Author Abstract
Daniel Frasco Nuclear Magnetic Resonance (NMR) is an important tool for measuring
Thermo Fisher Scientific hydrogen content in fuel. However, time domain NMR instruments
Madison, WI, USA (described below) offer little flexibility, and high-field frequency domain NMR
instruments (described below) can be prohibitively expensive. The Thermo
Application benefit Scientific™ picoSpin™ 80 Series II NMR spectrometer can serve as a flexible
The picoSpin 80 Series II NMR and economical alternative to time-domain and high-field NMR for this
spectrometer serves as a cost- application. In this note, a range of reference compounds and fuel standards
effective analytical tool for were analyzed using a picoSpin 80 Series II NMR spectrometer and the
hydrogen content determination. results were compared to their theoretical values as well as to those obtained
The picoSpin NMR spectrometer experimentally using a high-field NMR. For hydrogen content determination,
has the advantage of increased the picoSpin 80 Series II NMR spectrometer offers accuracy and precision
functionality compared to time- comparable to the high-field NMR measurements described in the literature.
domain instruments and a low
cost of ownership compared to Introduction
high-field NMR spectrometers. The combustion of hydrocarbon containing fossil fuels currently serves as
the largest source of energy worldwide. The more hydrogen attached to
Keywords each carbon in the hydrocarbon, the lower the oxidation state of that carbon
NMR, fuels, petroleum, hydrogen and the greater the energy released during combustion. Because of this, the
content hydrogen content is an important parameter of petroleum distillate products.1

A number of methods have been developed to determine the hydrogen content

in petroleum products. ASTM D5291, a combustion method for determining
the hydrogen content of petroleum fuels is limited in that it is not suitable for
samples with a low boiling range and is also a destructive technique which
restricts repeatability studies.2 Nuclear Magnetic Resonance (NMR) has long
been used for the determination of hydrogen content in petroleum hydrocarbons
and various ASTM methods have been published. Earlier NMR methods
(ASTM D3701 and ASTM D4808) made use of obsolete continuous-wave
spectrometers.3-4 Subsequently, an updated method ASTM D7171 was published
for determining the hydrogen content of middle distillates using a pulsed time-
The Thermo Scientific picoSpin 80 Series II
domain (TD) NMR spectrometer.5 TD-NMR quantifies materials using their
spectrometer. different relaxation properties and is found in quality control labs in a variety of
industries. However, it is inadequate for obtaining frequency spectra, thus is
not capable of structural analysis like frequency-domain NMR spectrometers.
Methods for determining hydrogen content in petroleum products have been
developed for high-field NMR spectrometers as well, but these instruments are
expensive to purchase and maintain, and require a specific NMR facility apart
from the factory or testing laboratory.1,6
In this application note, the picoSpin 80 Series II NMR Results and discussion
spectrometer was used to determine the hydrogen To determine hydrogen content using the pulsed picoSpin
content in a range of reference materials and petroleum 80 NMR spectrometer, an appropriate internal standard
fuel standards. The data shows that the picoSpin NMR was first selected. Hexamethyldisiloxane (HMDSO) was
spectrometer can serve as an effective, low-cost alternative chosen for this application because it is chemically inert,
to high-field NMR for hydrogen content determination. has a high boiling point, has similar hydrogen content
to relevant petroleum samples, and its 1H NMR signal
Experimental resonates in a region that does not overlap with the
A picoSpin 80 Series II benchtop NMR spectrometer samples of interest.1 Once the spectrum of the mixture
was used to acquire spectra for hydrogen content containing the internal standard HMDSO and the sample
determination. The spectrometer is an 82 MHz, pulsed, of interest was obtained, the following equation was used
Fourier transform 1H NMR. The instrument contains a 2 to determine the hydrogen content of the sample.
Tesla temperature controlled permanent magnet heated
to a temperature of 36 °C and is fitted with a 40 microliter
capillary cartridge used for sample introduction into the
spectrometer. All samples were purchased from Sigma In this equation, the percent hydrogen by weight was
Aldrich® and were used as received. determined using the recorded weight of the HMDSO
and sample as well as the integration of the peak area of
For hydrogen content determination, a vial was placed on the HMDSO singlet and the combined peak areas of the
an analytical balance and ~45mg of hexamethyldisiloxane sample signals. The percent hydrogen of HMSDO has
(HMDSO) was added. The mass was recorded and the a known value of 11.17% and was used in the equation.
balance was tared. Then ~300mg of the sample was An example of a 1H NMR spectrum used for calculating
added to the same vial and the mass was recorded. percent hydrogen content in this study is shown with the
The vial was capped and shaken until adequately mixed. integrations labeled in Figure 1.
Five runs were prepared for each sample. The samples
were injected into the capillary cartridge using 1 mL slip-
tip polypropylene syringes and 22 gauge blunt-tipped
needles. All spectra were acquired using 16 scans, 4000
acquisition points and a recycle delay of 40 seconds.

Figure 1: 1H NMR spectrum of diethyl malonate

with HMDSO internal standard
To examine the applicability of the picoSpin 80 NMR These relative error values generally track those reported
spectrometer for determining hydrogen content over a using a high-field NMR spectrometer, which examined 25
range of samples, several reference compounds with reference samples and reported a range of relative error
known hydrogen content, and published results from between -10.58% and 2.65%.1 On average, the picoSpin
other NMR instruments were analyzed.1,5,6 The reference method gave an average of 1.58% compared to the high-
compounds were first used to examine the precision of the field instrument which reported 2.75% relative error. The
hydrogen content measurement method. The six reference results indicate the low-field picoSpin NMR spectrometer
compounds selected represent a hydrogen content range was able to measure hydrogen content of selected reference
from ~7-15 percent (Table 1). The reported “Ave %H” is compounds with similar accuracy to a high-field instrument.1
an average of 5 runs and the percent relative standard
deviation (%RSD) for each compound is shown. In all Three ASTM fuel standards were then analyzed and
cases, the %RSD was below 1%. The average %RSD of results are summarized in Table 3.
the six compounds analyzed in this work was 0.44%, on
par with the reported results using a high-field instrument.1

The accuracy of this method for determining hydrogen

content was also examined by using the theoretical %
hydrogen content to calculate the relative percent error
for each sample. The 6 reference samples examined with
the picoSpin spectrometer gave a range of relative error
between 0.56% and 2.59%.

Sample Ave %H %RSD Sample Ave %H %H theoretical %Relative error

Diethyl Malonate 7.72 0.36 Diethyl Malonate 7.72 7.55 2.22

Toluene 8.80 0.32 Toluene 8.80 8.75 0.56

Mesitylene 10.23 0.19 Mesitylene 10.23 10.06 1.68

Cyclohexanone 10.44 0.44 Cyclohexanone 10.44 10.27 1.63

2-Nonanone 13.09 0.41 2-Nonanone 13.09 12.76 2.59

Dodecane 15.51 0.89 Dodecane 15.51 15.39 0.77

Table 1: Average hydrogen content and % relative standard deviation Table 2: Accuracy of picoSpin 80 NMR spectrometer for determining
of five runs per reference sample determined using picoSpin 80 NMR. hydrogen content using five runs per reference sample.

Standard D5307 %H by NMR Standard D5580 %H by NMR Standard D5134 %H by NMR

Test 1 15.38 Test 1 14.36 Test 1 14.33

Test 2 15.31 Test 2 14.37 Test 2 14.38

Test 3 15.32 Test 3 14.60 Test 3 14.27

Test 4 15.23 Test 4 14.30 Test 4 14.14

Test 5 15.41 Test 5 14.38 Test 5 14.50

Average 15.33 Average 14.40 Average 14.32

Theoretical 15.18 Theoretical 14.61 Theoretical 14.38

%Relative Error 0.98 %Relative Error -1.43 %Relative Error -0.39

%RSD 0.41 %RSD 0.71 %RSD 0.83

Table 3: . Hydrogen content determination using fuel standards.

ASTM D5307 Crude Oil Internal Standard has a calculated References
theoretical hydrogen content of 15.18% and is made up of 1. Mondal, S.; Kumar, R.; Bansal, V.; et al. J. Anal. Sci.
a mixture of C14-C17 hydrocarbons. This is comparable to Technol. 2015, 6: 24
diesel fuel which typically contains a mixture of C10-C19
hydrocarbons. The average percent hydrogen using the 2. ASTM-D5291, Standard Test Methods for Instrumental
picoSpin spectrometer was determined to be 15.33% Determination of Carbon, Hydrogen, and Nitrogen
which gave a relative error of 0.98% when compared to in Petroleum Products and Lubricants. ASTM: West
the theoretical value. The ASTM D5580 Calibration Mix Conshohocken, PA, 1992
4 is composed primarily of 2,2,4-trimethylpentane, an
important component in gasoline. The picoSpin method 3. ASTM-D3701, Standard Test Method for Hydrogen
yielded a hydrogen content of 14.40%, which gave a relative Content of Aviation Fuels by Low Resolution Nuclear
error of -1.43%. The final standard analyzed was ASTM Magnetic Resonance Spectroscopy. ASTM: West
D5134 Splitter Linearity Check Mix. The obtained hydrogen Conshohocken, PA, 1992
content of 14.32% also matched closely with its theoretical
value of 14.38% with a relative error of -0.39%. ASTM 4. ASTM-D4808, Standard Test Methods for Hydrogen
D5580 contained a mixture of C6-C8 hydrocarbons, and Content of Light Distillates, Middle Distillates, Gas
ASTM 5134 contains a mixture of C6-C9 hydrocarbons, Oils, and Residua by Low Resolution Nuclear
both of which are within the typical C4-C12 hydrocarbon Magnetic Resonance Spectroscopy. ASTM: West
range of gasoline. Conshohocken, PA, 1992

Conclusions 5. ASTM-D7171, Standard Test Method for Hydrogen

Although there are existing ASTM methods for determining Content of Middle Distillate Petroleum Products
the hydrogen content of fuel and petroleum using NMR Low-Resolution Pulsed Nuclear Magnetic Resonance
spectroscopy, the instruments involved are either obsolete, Spectroscopy. ASTM: West Conshohocken, PA, 2011
or limited in their applications. It has been previously
shown that high-field NMR is capable of determining 6. Khadim, M. A.; Wolny, R. A., Al-Dhuwaihi, A. S.; et. al.
the hydrogen content of fuel samples.1,6 In this work, six Arab. J. Sci. Eng. 2003, 28(2A) 147-162
reference compounds representing a range of hydrogen
contents were analyzed by the picoSpin 80 benchtop
NMR spectrometer. The accuracy and precision of the
measurement were found to be similar to those obtained
with high field instruments.

Three fuel standards mimicking actual fuels were also

analyzed, and the hydrogen content agree well with the
theoretical and high-field NMR measurements. These results
indicate that the picoSpin 80 spectrometer is capable of
providing hydrogen content information with similar accuracy
and precision to high-field instruments NMR at a significantly
reduced cost of ownership.

Find out more at thermofisher.com/nmr

©2017 Thermo Fisher Scientific Inc. All rights reserved. Sigma-Aldrich is a registered trademark of Sigma-
Aldrich Co. All other trademarks are the property of Thermo Fisher Scientific and its subsidiaries unless
otherwise specified. AN52979_E_10/17M
APPLICATION NOTE AN52995

Evaluation of fluoroorganic compounds with

benchtop 19F NMR
Author Abstract
Dean Antic, Ph.D., Fluorine-19 NMR adds a new dimension in the analysis portfolio of fluorine
Thermo Fisher Scientific containing compounds. The broader chemical shift range in 19F NMR helps
Boulder, CO, USA resolve individual fluorine containing functional groups, while the often large
variable magnitude of 19F-19F and 1H-19F coupling provides additional insight
Application benefits into structural effects. Furthermore, first-order coupling and highly resolved
With a dedicated Thermo Scientific™ resonance lines simplify analysis.
picoSpin™ 80 19F nuclear magnetic
resonance (NMR) spectrometer, Introduction
spectral acquisition is as easy as Fluorine (19F) plays an important role in pharmaceutical, agrochemical, and
proton NMR. Due to the unique medicinal chemistry because the judicious placement of fluorine atoms in
attributes of fluorine nuclei, the 19F a molecule can have a significant influence on its chemical and physical
NMR spectra are easy to interpret, properties. While the steric impact of replacing a hydrogen (1H) atom with
and provide a wealth of molecular fluorine is minimal, the electron withdrawing and inductive field effects
structure information as well as its from a single fluorine nucleus are profound, affecting such properties as
associated chemical environment. acidity, lipophilicity, and polarity. Examples of fluoroorganic compounds,
with varying degree of structural complexity and fluorination, are shown in
Keywords Figure 1. Biologically active pharmaceutical and agrochemical compounds
Fluoroorganic compounds, 19F NMR, incorporating fluorine are typically “lightly” fluorinated, containing one or only
fluorotelomer, hydrofluorocarbon, a few fluorine-containing substituents. Electronic liquids, on the other hand,
J coupling, multiplicity are oligo perfluorocarbon compounds. They are chemically inert and are
often used as lubricants and in low temperature heat transfer applications.

Figure 1: Chemical structures of the fluoroorganic compounds

used in this study. (1) Fluorobenzene; (2) 4-fluorobenzyl bromide; (3)
[4-(trifluoromethyl)phenyl]methanamine; (4) perfluoro-n-butyl ethylene,
trade name Zonyl® PFBE; and (5) 2H,3H-decafluoropentane, trade name
Vertrel® XF.
 F is one of the most studied nuclides after 1H and 13C in
19
fluorotrichloromethane (0 ppm), trifluoromethylbenzene
NMR spectroscopy. The advantages of 19F NMR are multifold: (-63.2 ppm), trifluoroacetic acid (-76.2 ppm), and ethyl
trifluoroacetate (-75.8 ppm), are also used to capture the
1. For lightly fluorinated bioactive molecules, the relatively full range of possible chemical shifts.
small number of fluorine atoms yield fewer signals for
easy spectral interpretation. Spectra were acquired using the following acquisition
parameters: a 90° RF excitation pulse, a 1000 ms
2. Fluorine couples strongly to other 19F nuclides as far acquisition time, and a 10 second recycle delay. The
as six bonds away, while also coupling to proximate spectral width was adjusted to capture the large chemical
1
H nuclei and carbon atoms. The resulting 19F NMR shift dispersion of the 19F spectrum. All spectra were
spectrum is therefore information-rich in molecular acquired with signal averaging. Spectral data were stored
structure as well as associated chemical environment. in a JCAMP-DX file format and imported into MestreLab
Research’s Mnova™ NMR analysis program for processing.
3. The 19F NMR spectra are typically first-order in nature in Standardized data processing was applied across all
that spin-spin coupling follows the n+1 rule for multiplicity. spectra, specifically, by applying zero filling, applying phase
correction, and filtering using exponential Apodization.
4. Being the most electronegative element on the periodic
table, 19F displays large shift dispersion (from -300 ppm Results and discussion
to 400 ppm), resulting in far larger chemical shifts in 19F Examples of lightly fluorinated compounds
spectra than in 1H, and a much smaller probability of The 19F NMR spectra of two mono-substituted aromatic
peak overlapping. fluoroorganic compounds, and one trifluoromethyl-
substituted phenylmethylamine are presented in Figure 2.
To demonstrate the wealth of 19F spectral information, in Fluorine substituents on the aromatic ring generally
this application note, we present the 19F NMR spectra of a absorb in the same region (from -200 ppm to -100 ppm)
series of commercially available fluoroorganic compounds, because shielding zones due to the ring currents in the
obtained using a Thermo Fisher Scientific picoSpin 80 19F benzene ring have little influence on fluorine atoms. The
NMR spectrometer. mono-fluorine substituted aromatic experiences 3 JFH and
4
JFH coupling to ring protons, resulting in complicated but
Experimental well-defined multiplets. The trifluoromethyl group (CF3) in
Spectra were acquired using a picoSpin 80 NMR compound 3, on the other hand, appears as a singlet.
spectrometer. The spectrometer is a pulsed, Fourier
transform 19F NMR permanent magnet instrument
equipped with a capillary cartridge probe. To ensure the
highest sensitivity, the spectrometer was tuned to the
fluorine Larmor frequency of 77 MHz. The spectrometer’s
capillary cartridge was fitted with micro-fluidic inlet and
outlet connectors that allow liquid sample injection into
the spectrometer’s RF coil. The fluid path was Teflon/
Quartz capillary tubing with a total flowpath volume of
40 microliters (μL). Liquid samples were introduced by
manual injection using a disposable 1 mL syringe and
a 22-gauge blunt tip needle. Hexafluorobenzene (HFB,
C6F6) was added to fluorinated samples at approximately
1% (v/v) concentration as an internal shift reference.
Hexafluorobenzene is a high fluorine density compound
providing a strong fluorine signal, and is assigned a
chemical shift value of -164 ppm. Due to large shift
Figure 2: Full 19F NMR spectra of, from top down, fluorobenzene
dispersion and solubility requirements of fluorine (compound 1), 4-fluorobenzyl bromide (compound 2), and
compounds, other reference compounds, such as [4-(trifluoromethyl)phenyl]methanamine (compound 3).
Note that J coupling is indirect, through-bond scalar The insets show the J-coupling trees for the peaks
coupling between nuclei of like spin. The superscripts of well-defined multiplicity (peaks 4, 5, and 6). The
preceding the J term, e.g., 2J, 3 J, or 4 J, indicate the number magnitude of the 3 JFF coupling constants are of the
of bonds separating the coupled nuclei. 3 J coupling is expected size for linear, symmetric primary and
most commonly observed in 1H NMR, whereas with 19F secondary fluoroalkyl groups. For peak 3, the coupling
NMR, coupling 2-6 bonds away is routinely observed. The of fluorine atoms to adjacent protons (3 JFH) at position C3
subscripts, e.g., JHH, JFH, or JFF, indicate which nuclei are disrupts the symmetry at this position, causing the loss of
coupled. Coupling constants provide valuable information a well-defined multiplet.
on which nuclei are close to one another. Coupled nuclei
split the signal intensity of an otherwise uncoupled nuclei, An example of a hydrofluorocarbon
resulting in a multiplet signal. The multiplicity, following the Hydrofluorocarbon (HFC) electronic fluids are used in
n+1 rule for first-order coupling, signifies the number of a variety of applications, such as heat transfer fluids,
adjacent coupled nuclei, n. lubricants, and in precision and optics cleaning. Figures 4
and 5 are the 1H and 19F spectra, respectively of 2H,3H-
An example of a fluorotelomer decafluoropentane, a specialty fluid sold by DuPont™
Fluorotelomers are fluorocarbon-based oligomers under the trade name Vertrel® XF.
synthesized by radical polymerization. They are used
in a variety of manufacturing processes as flame In the 1H spectrum (Figure 4), in addition to the H-H
retardants, and as non-conductive coatings due to their coupling 3 JHH, both protons experience strong 1H-19F
lipophobicity. Fluorotelomers are also the basis of many geminal coupling (2JFH; coupling of spin ½ nuclei attached
environmentally persistent perfluorinated carboxylic acids to the same carbon center) to fluorine, and vicinal
because of their use as surfactants. coupling (3 JFH; coupling of spin ½ nuclei on adjacent
carbon centers) to adjacent fluorine atoms, yielding
The compound perfluoro-n-butyl ethylene is a a complex multiplet splitting pattern. The complexity
fluorotelomer intermediate sold under the trade name is further compounded by the asymmetry of proton
Zonyl® PFBE. It has an ethylene functional group and substitution at the C2 and C3 positions and slightly
fluorocarbon backbone. The higher spectral resolution of different chemical shifts for the two protons. Multiplet
the fluorine peaks makes multiplet analysis and structural analysis of the splitting pattern suggests a dddd class,
determination facile (Figure 3). with a 2JFH coupling constant of 43.4 Hz.

Figure 3: Full 19F NMR spectrum of perfluoro-n-butyl ethylene. Inset: Figure 4: Full 1H NMR spectrum of 2H,3H-decafluoropentane (C5H2F10;
Multiplet analysis of the splitting pattern of individual peaks, with a neat) with TMS added as a chemical shift reference. Multiplet analysis
J-coupling tree and coupling constants overlaid on the spectrum. of the splitting pattern reveals a dddd class; a J-coupling tree and
coupling constants are overlaid on the spectrum.
In the 19F spectrum (Figure 5), primary and secondary Conclusions
alkyl fluorides experience vastly different shielding along Fluorine-19 NMR compliments 1H and 13C NMR in
the carbon backbone, spanning a range from -220 ppm structure determination. The broader chemical shift range
to -75 ppm, allowing for the analysis of individual fluorine in 19F NMR helps resolve individual fluorine containing
groups separately. Confirmation of the strong geminal 2JFH functional groups, while the often-large variable magnitude
coupling observed in the 1H spectrum is in the multiplet of 19F-19F and 1H-19F coupling provides additional insight
analysis of the splitting pattern of fluorine atoms at into structural effects. First-order coupling and highly
positions C2 and C3 (inset for peaks 2 and 3). resolved resonance lines also simplify analysis. Overall,
19
F NMR adds a new dimension in the analysis portfolio of
fluorine containing compounds.

Figure 5: Full 19F NMR spectrum of 2H,3H-decafluoropentane (C5H2F10;

neat) with C6F6 added as a chemical shift reference. Inset: Expanded
view of chemical shift regions showing complex, multiplet splitting
patterns arising from 2 JFH and 3 JFH coupling, and molecular asymmetry.
J-coupling trees and coupling constants are overlaid on the spectrum.

Find out more at www.thermofisher.com/NMR

©2017 Thermo Fisher Scientific Inc. All rights reserved. Mnova is a trademark of MestreLab Research, Inc. All
other trademarks are the property of Thermo Fisher Scientific and its subsidiaries unless otherwise specified.
AN52995_E 10/17M
APPLICATION NOTE AN52301

Classification of polyethylene by Raman spectroscopy

Author Application benefits

Mohammed Ibrahim, Ph.D., Raman spectroscopy provides molecular level structural information, and
Herman He, Ph.D., is advantageous in classifying polyethylene (PE). Sample preparation is
Thermo Fisher Scientific, USA minimal. The method is non-destructive, and the analysis is fast (once the
TQ methods are established).
Keywords
Polyethylene (PE), density, Raman Introduction
Spectroscopy, Raman Microscopy, Polyethylene (PE) is one of the most common plastics in the world with
classification, Discriminant Analysis annual global production of around 80 million tons.1 Based on density,
polyethylene is generally classified as high-density polyethylene (HDPE, >
0.940 g/cm3) or low-density polyethylene (LDPE, < 0.930 g/cm3).2 These
different density polyethylene’s have vastly different physical, chemical, and
mechanical properties, and hence are used in different applications. For
example, HDPE is primarily used for milk jugs, detergent bottles, garbage
containers, and water pipes, due to its high tensile strength; LDPE, on the
other hand, has a lower tensile strength and is used mainly for plastic bags
and wraps. Therefore, density is one the most important properties of
polyethylene, and classifying them according to their density is essential for
proper PE specification.

Bulk PEs are manufactured as pellets (resins, granules), and later converted to
other forms (such as films and pipes) using extrusion or molding processes.
They are also made into multilayer films for a wide range of industrial
applications like food and consumer product packaging. The density of bulk
PE pellets and single-layer PE films can be measured and classified with
relative ease using several standard techniques: ISO 1183-1/ASTM D792
(immersion method),3 ISO 1183-2/ASTM D1505 (density gradient method),4
and ASTM D4883 (ultrasound method).5 However, all these techniques
require the PE in its “pure” form, which can be challenging in the case of PE
in multilayer films. Extensive sample preparations (microtoming, separation
of layers by dissolving in solvents) are often required6 to isolate the PE layer
before analysis, which can be labor-intensive and time-consuming.
Raman spectroscopy is sensitive to changes in the A 532 nm laser was used with a 2 mW laser power at the
molecular structure level of PE, such as the degree of sample. A 10x objective and a 50 μm slit aperture were used
crystallinity, which is the key determining factor of PE to obtain more representative spectra from the samples.
density.7,8 More importantly, the confocal capability of Total acquisition time for each spectrum was 30 seconds
Raman microscopy allows for facile in situ analysis of (3 second exposure x 10 exposures). Thermo Scientific™
individual PE layers in multilayer films without the need to OMNIC™ software was used for operation of the DXR2
isolate the PE layer. To our best knowledge, PE density Raman Microscope, and collection of Raman spectra;
measurement using Raman has been limited to PE Thermo Scientific™ TQ Analyst™ software was used for
pellets.7,8 In this work, we want to systematically explore chemometric analysis of the Raman data.
the feasibility of using confocal Raman microscopy for PE
film density analysis, both qualitatively and quantitatively. Results and discussion
We demonstrate that Raman microscopy in combination Raman spectra
with the discriminant analysis method can be successfully Representative Raman spectra of HDPE and LDPE
applied to distinguish HDPE and LDPE in both pellet samples, in both pellet and film forms, are shown in
and film forms. In a subsequent application note, we will Figure 1. There are noticeable differences between HDPE
detail the quantitative determination of PE density using a and LDPE spectra, for both pellets and films. In the CH2
confocal Raman microscope. bending and the CH2 twisting region, the intensity of
the CH2 bending mode at 1416 cm-1 (relative to the CH2
Experimental bending mode at 1440 cm-1) is higher for HDPE than for
Sample description LDPE. This observation agrees with the previous reports
A total of 16 PE samples (10 pellets and 6 films) with that the 1416 cm-1 and 1440 cm-1 peaks are indicators of
known densities were used for the classification studies. crystalline and amorphous PE phases, respectively.7-10
All samples were used as received. The higher the degree of crystallinity, the higher the
density. The differences between HDPE and LDPE
Method conditions are also pronounced in the C-H stretching region. The
A Thermo Scientific™ DXR2™ Raman Microscope was intensity of the symmetric CH2 stretching mode at 2848
used for the collection of Raman data. For each type/class cm-1 (relative to the asymmetric CH2 stretching mode at
of the pellet samples, Raman spectra were collected from 2882 cm-1) appears to be higher for LDPE compared to
3 different pellets and averaged. For each film sample, HDPE. Since the C-H stretching (2825-2970 cm-1) and the
Raman spectra were collected from 3-4 locations across CH2 bending regions (1398-1470 cm-1) are sensitive to the
the surface of the sample. An averaged spectrum was density of PE, these regions were selected for subsequent
then used for final analysis. discriminant analysis.

Figure 1: Representative Raman spectra of HDPE and LDPE pellets and films. (A) Full spectral range. (B) C-H stretching region. (C) CH2 bending and
CH2 twisting region.
Data processing
The raw Raman spectra were processed using Norris Norris derivative is effective in removing background drift
2nd derivative, and the resulting spectra were further in Raman spectra caused by fluorescence, whereas SNV
processed by standard normal variate (SNV). Examples of is effective in compensating such variations as sample
the data processing are shown in Figure 2. surfaces and laser penetration depths.11-12

Figure 2: Norris 2nd Derivative and SNV processed sample spectra. (A) Full spectral range. (B) C-H stretching region. (C) CH2 bending region. HDPE
spectra are in red in both B and C Plots.

Classification of PE by discriminant analysis

Calibration results
PE sample Actual class Usage Calculated class Distance to HDPE Distance to LDPE
1 HDPE Calibration HDPE 0.89 4.23
2 HDPE Calibration HDPE 0.77 4.97

3 HDPE Calibration HDPE 0.67 4.05

4 HDPE Validation HDPE 1.33 3.69

5 HDPE
Pellets

Calibration HDPE 0.68 5.14

6 HDPE Calibration HDPE 0.76 4.91

7 LDPE Calibration LDPE 4.01 0.99

8 LDPE Calibration LDPE 5.32 0.78

9 LDPE Validation LDPE 3.79 0.94

10 LDPE Calibration LDPE 4.71 0.77

11 HDPE Validation HDPE 1.23 4.79

12 HDPE Calibration HDPE 1.24 4.66

13 HDPE Calibration HDPE 1.30 4.93

Films

14 LDPE Validation LDPE 4.97 0.84

15 LDPE Calibration LDPE 4.84 0.57

16 LDPE Calibration LDPE 4.54 1.19

Table 1: PE sample class types and calibration results

The discriminant analysis classification method with its own class vs. the M distance to the other class. The
principal component analysis (PCA) algorithm13 from the average M distance for each sample to its own class
TQ Analyst software package was applied to distinguish was about 1, but the average distance to the other class
HDPE vs. LDPE. A total of 12 samples, a mix of pellets is over 4, as listed in Table 1. For both sample classes
and films, were used as the calibration standards. Four (HDPE and LDPE), there is no separation between pellets
additional samples (one HDPE pellet, one HDPE film, one and films, suggesting the sample form (pellet vs. film) has
LDPE pellet, and one LDPE film) were selected as the little, if any, impact on the methodology.
validation standards (Table 1).

PCA derives the principal components (PC) or the

significant spectral information from the spectral variance
of the calibration sample set. The number of significant
PCs represents the number of independent variables
affecting spectral responses, including but not limited to:
concentration, impurities, opaqueness, and sample color.
Scores of PCs depict the projected sample spectrum
in the principal component domain. The PCA-based
classification method then calculates the Mahalanobis (M)
distance, defined by the distance between the sample and
the center of each cluster in the PC domain. The sample
is classified as belonging to a class if M < 3, and rejected
from a class if M > 3.

Number of PCs Cumulative Variance%

0 0
1 80.96
2 98.03
3 98.83
4 99.46
5 99.74

Table 2: Impact of number of PCs on variance coverage Figure 3: 3-D cluster plot of HDPE and LDPE samples. The • are the
calibration samples, and the + are the validation samples.

The number of PCs has a direct impact on the robustness

of the discriminant analysis. As shown in Table 2, the first
several principal components represent the majority of
the spectral variation. Five PCA factors, accounting for
99.7% of the total spectral variance, were used in this
classification model.

Figure 3 is a 3-D plot of PCA Clusters of HDPE and

LDPE. The HDPE samples (red) and LDPE (blue)
samples are located on the opposite ends of the cube.
Notwithstanding PC1 accounting for ~81% of the total
spectral variation in the data set, the separation between
Figure 4: TQ Analyst software discriminant analysis calibration output
HDPE and LDPE is mainly in the PC2 dimension. In the for PEs with different densities. The two different types of PEs: HDPE
current case, the PC2 dimension seems to be closely and LDPE are clearly separated, and correctly classified. The calibration
results are also shown in Table 1.
correlated to the PE density. Figure 4 shows the cross-
validation results of the 16 samples, the M distance to
The Raman spectrum of a previously unused pellet sample The sample was successfully classified as HDPE with an
with known density was used to test the established M value of 0.68.
discriminant method, and the result is shown in Figure 5.

Figure 5: Classification of a PE sample by using its Raman spectrum and the discriminant method created by the TQ Analyst software.

Conclusions
In this application note, we have successfully
demonstrated the use of a Thermo Scientific DXR2 Raman
Microscope, in combination with the TQ Analyst software,
to classify polyethylene’s of different density classes in
both pellet and film forms. Raman spectroscopy is non-
destructive and requires minimal sample preparation.
The classification method was created solely based on
the Raman spectral features of LDPE and HDPE and
was indifferent to the sample forms. Once the method is
established, PE samples, pellets or films, can be correctly
classified within minutes. Moreover, this work expands
the scope of the previously reported study on PE pellets
to include PE films, which broadens its applicability in the
Thermo Scientific DXR2 Raman Microscope
plastic/polymer industry as well as many downstream
industries. The described methodology should be
applicable for in situ classification of thin PE layer(s) in
multilayer films.
References
1. Piringer O.G. and Baner A.L., ed. Plastic Packaging: Interactions with Food and Pharmaceuticals. 2nd ed. Weinheim:
Wiley-VCH; 2008.
2. Polyethylene, The Essential Chemical Industry - online,
http://www.essentialchemicalindustry.org/polymers/polyethene.html, retrieved on 9/10/2017.
3. (a) ISO 1183-1:2012, Plastics − Methods for determining the density of non-cellular plastics − Part 1: Immersion
method, liquid pyknometer method and titration method; ANSI New York, NY. (b) ASTM D792-13, Standard test
methods for density and specific gravity (relative density) of plastics by displacement; ASTM International, West
Conshohocken, PA, 2013.
4. (a) ISO 1183-2:2004, Plastics — Methods for determining the density of non-cellular plastics — Part 2: Density
gradient column method. (b)ASTM D1505-10, Standard test method for density of plastics by the density-gradient
technique; ASTM International, West Conshohocken, PA, 2010.
5. ASTM D4883-08, Standard test method for density of polyethylene by the ultrasound technique; ASTM International,
West Conshohocken, PA, 2008.
6. Mieth A., Hoekstra E., and Simoneau C. Guidance for the identification of polymers in multilayer films used in food
contact materials: User guide of selected practices to determine the nature of layers, EUR 27816 EN, 2016; doi:
10.2788/10593.
7. Strobl G.R. and Hagedorn W. Raman spectroscopic method for determining the crystallinity of polyethylene, J.
Polym. Sci. B Polym. Phys., 1978, 16, 1181-1193.
8. Sato H., Shimoyama M., Kamiya T., Amari T., Sasic S., Ninomiya T., Siesler H.W., and Ozaki Y. Raman spectra of
high-density, low-density, and linear low-density polyethylene pellets and prediction of their physical properties by
multivariate data analysis, J. Appl. Polym. Sci., 2002, 86, 443–448.
9. Migler K.B., Kotula A.P., Hight Walker A.R. Trans-rich structures in early stage crystallization of polyethylene,
Macromolecules, 2015, 48, 4555−4561.
10. Kida T., Hiejima Y., and Nitta K-H. Raman spectroscopic study of high-density polyethylene during tensile
deformation, Int. J. Exp. Spectroscopic Tech., 2016, 1:001.
11. Simone, E., Saleemi A.N., and Nagy Z.K. Application of quantitative Raman spectroscopy for the monitoring of
polymorphic transformation in crystallization processes using a good calibration practice procedure, Chem. Eng.
Res. Des. 2014, 92, 594–611.
12. Huang J., Romero-Torres S., and Moshgbar M. Practical considerations in data pre-treatment for NIR and Raman
spectroscopy, Am. Pharm. Rev. 2010, 13 (6).
13. Thermo Scientific Product Overview. TQ Analyst Software Chemometric Algorithms, 2009.

Find out more at www.thermofisher.com/Raman

©2017 Thermo Fisher Scientific Inc. All trademarks are the property of Thermo Fisher Scientific and its
subsidiaries unless otherwise specified. AN52301_E 11/17M
APPLICATION NOTE AN52304

FT-Raman: an invaluable addition to the forensic arsenal to

combat the opioid epidemic
Authors Introduction
Suja Sukumaran, Ph.D.1 Illicitly manufactured fentanyl is becoming a major driver of opioid overdose.
Rui Chen, Ph.D.1 According to the Centers for Disease Control and Prevention (CDC), illegal,
Shea A Schleman2 lab-made fentanyl was involved in more than 50% of opioid overdose deaths
Laura J Bell2 in 2016.1 The growing opioid epidemic presents multifaceted challenges for
1
Thermo Fisher Scientific, USA law enforcement, first responders, and forensic lab personnel. In particular, the
2
Albuquerque Police Department, high potency of fentanyl and fentanyl analogues makes accidental exposure
Albuquerque, New Mexico life-threatening, with ingestion, inhalation, and absorption through the skin as
possible exposure routes. With a lethal dose of only a few milligrams, fentanyl
is considered 50-100 times more potent than morphine. Carfentanil, a fentanyl
analogue, is approximately 100 times more potent than fentanyl. Consequently,
the CDC has issued a health alert on the rise of unintentional overdoses of
clandestinely produced and trafficked fentanyl in the form of counterfeit pills and
heroin adulterants.1-2 The situation is further exacerbated by the waves of new
synthetic opiates. Through chemical modification, new potent fentanyl analogues
are created at a fast pace in underground labs, keeping authorities constantly
on their toes to identify these emerging chemical entities and to understand their
pharmacology and toxicology.

The challenges in the detection, identification, and screening of

synthetic opioids mandate the use of multiple analytical techniques
and instrumentation, both field-deployable and laboratory-based, in a
concerted and holistic manner. For example, The CDC advises two-tiered
testing, an enzyme-linked immunosorbent assay (ELISA) screen followed
by gas chromatography/mass spectrometry (GC/MS) to identify the fentanyl
compound in blood and urine in the cases of suspected overdose. In the
meantime, liquid chromatography / mass spectrometry (LC/MS), LC/MS/
MS, ion mobility spectrometry (IMS), and thermal desorption direct analysis
in real time mass spectrometry (TD-DART-MS) are being explored to meet
the demand on low detection limit for many case samples where fentanyl is
present with other drugs and cutting agents at low concentration.2
Raman spectroscopy has long been used for the Results and discussion
detection and identification of illicit drugs and adulterants, Figure 2A shows the Raman spectrum of bulk fentanyl
and offers a valuable addition to the forensic toolbox in a double bag, directly acquired without any sample
for the analysis of fentanyl and fentanyl analogues. manipulation. Since the laser spot size is approximately
FT-Raman utilizes a long-wavelength laser (1064 nm) 60 μm, small quantity of samples, such as a few granules,
which greatly reduces fluorescence and produces high can be analyzed with ease. An example is shown in
signal-to-noise spectra, making it well suited for many Figure 2B. While Figure 2B exhibits a slightly higher
narcotic samples that fluoresce. It also enables sampling noise level compared to Figure 2A, both samples were
through glass vials, polymer blister packs, and plastic nonetheless positively identified as fentanyl citrate through
evidence bags; hence, often requires little to no sample library search.
preparation. It is nondestructive and allows the custody
chain to be maintained, given the possibility of qualitative
and quantitative evaluation of the sample to confirm its
integrity from its seizure until the sample is placed in the
evidence file.3

In this application note, analyses of fentanyl as well as

other illicit drugs using FT-Raman spectroscopy are
presented. The advantages of FT-Raman for the detection Figure 2: Raman spectra of (A) seized bulk fentanyl in a double-bag;
and identification of illicit drugs are also discussed. (B) a few fentanyl granules in a double-bag; and (C) fentanyl citrate
from DEA Raman library.

Materials and methods Raman spectroscopy is sensitive to both chemical and

All samples analyzed in this study were prepared at physical properties. Its unique selection rules generate a
Albuquerque Police department. Samples were analyzed molecular fingerprint that is well suited to the differentiation
in double bags, blister packs or glass vials, and handled in between many illicit drug compounds and their analogues.
compliance with Schedule I controlled substance handling Figure 3 shows the Raman spectra of a tablet as well
protocols. A Thermo Scientific™ Nicolet™ iS50 FT-IR as the 3,4-methylenedioxyamphetamine (MDA) and
spectrometer equipped with a calcium fluoride beam 3,4-methylenedioxymethamphetamine (MDMA) standards
splitter and an FT-Raman module with a 1064 nm laser from the library. Despite the minute structural difference
was used for all analyses (Figure 1). A total of 64 scans between MDA and MDMA (inset of Figure 3), Raman
were co-added for each spectrum at 8 cm-1 resolution, spectroscopy can unambiguously distinguish the two. In
with a total acquisition time of ~75s. The laser power this case, the tablet was identified as MDA with a match
was set at 0.5 mW. The spectra were searched against score of 83, as opposed to a match score of 66 for MDMA.
Thermo Scientific Law Enforcement and Security (LEnS)
Raman and DEA Raman libraries using OMNIC library
search and Specta multicomponent search options.

Figure 3: Raman spectra of (A) a tablet; (B) 3,4-methylenedioxyamphetamine

(MDA) standard from the library; and (C) 3,4-methylenedioxymethamphetamine
(MDMA) standard from the library.

Seized street drugs are often present in the form of

mixtures. Figure 4 shows the Raman spectrum of an
off-white powder sample in an evidence bag. Initial library
search showed similar match score for lactose and
cocaine, suggesting that the sample contains multiple
components. The spectrum was then subjected to the
OMNIC Specta multi-component search, and the sample
Figure 1: An iS50 Raman module for the Thermo Scientific Nicolet was identified as a mixture of cocaine and β-D-lactose, a
iS50 FT-IR spectrometer. commonly used cutting agent.
Conclusions References
The Nicolet iS50 FT-IR spectrometer equipped with the 1. https://www.cdc.gov/drugoverdose/opioids/fentanyl.html
iS50 Raman module offers an invaluable tool for the 2. Edward Sisco, Jennifer Verkouteren, Jessica
forensics labs to combat the growing opioid epidemic. Staymates, Jeffrey Lawrence, Forensic Chemistry,
Raman analysis requires little to no sample preparation Volume 4, June 2017, Pages 108-115,
and allows for direct measurements through glass 3. Ciro Augusto Fernandes de Oliveira Penido, Marcos
vials and evidence bags, greatly reducing the risk of Tadeu Tavares Pacheco, Igor K. Lednev, Landulfo
lab personnel’s accidental exposure to high potency Silveira Jr., Journal of Raman Spectoscopy, 47 (1),
drugs such as fentanyl and fentanyl analogues. While January 2016, Pages 28–38.
the measurement of bulk narcotic samples is fast
and straightforward, with a laser spot size of 60 μm,
acquisition of high quality Raman spectra from as little
as few granules is also possible. Combined with DEA
and LEnS Raman libraries, OMNIC Specta enables
identification of multi-component illicit drugs with
confidence. Raman spectroscopy using the iS50 Raman
module is a safe, fast, and complementary technique for
the screening, detection and identification of illicit drugs.

Figure 4: Raman spectra of a seized street drug sample containing cocaine and β-D-lactose.

Find out more at thermofisher.com/is50

©2017 Thermo Fisher Scientific Inc. All rights reserved. All trademarks are the property of Thermo Fisher
Scientific and its subsidiaries unless otherwise specified. AN52304_E_11/17M
 APPLICATION NOTE AN52326

Leveraging the lateral spatial resolution of a confocal

Raman microscope to resolve micron to sub-micron
layers in polymer laminates
Author Introduction
Mohammed Ibrahim, Ph.D. and Rui Polymer laminates are used in a wide range of industries from food to
Chen, Ph.D. pharmaceuticals. By adhering together multiple polymeric layers of different
Thermo Fisher Scientific, Madison, WI chemistry, physical properties, and thickness, desired chemical, mechanical
and barrier properties can be achieved to suit different packaging applications.
Keywords As the technological advancement and manufacturing sophistication continue
laminates, polymer, confocal to drive the production of more complex and thinner laminate structures, there
Raman microscope, spatial is an ever-increasing demand on laminate analysis for product quality control,
resolution failure analysis, and competitive reverse engineering.

Application benefits The analytical tools for laminate analysis include optical microscopy, differential
With a properly configured scanning calorimetry (DSC), Fourier Transform infrared spectroscopy (FT-IR),
Thermo Scientific™ DXR™2 Raman and Raman spectroscopy. Confocal Raman microscopy, in particular, offers
microscope, it was possible to resolve many advantages. Raman spectroscopy is sensitive to both chemical and
a thin polymer layer with a thickness physical properties. Its unique selection rules generate a molecular fingerprint
(≈0.4 μm) that’s very close to the that is well suited to material identification. Confocal Raman microscopy often
theoretical spatial resolution limit. utilizes short-wavelength visible and NIR (400 – 785 nm) lasers for sample
excitation, which renders increased sensitivity because Raman signal intensity
Thermo Fisher Scientific is inversely proportional to the 4th power of the laser wavelength. Furthermore,
solutions the spatial resolution is also inversely proportional to the laser wavelength. The
• DXR2 Raman microscope shorter the wavelength, the higher the resolution.1,2

• Thermo Scientific™ OMNIC™ for

In this application note, two cases of polymer laminate analysis using a
Dispersive Raman software suite
Thermo Scientific DXR2 Raman Microscope are presented. The implications of
instrument configurations, including laser wavelength, magnification objective,
and pinhole size, on lateral spatial resolution are discussed in details.
Materials and methods
Two polymer laminate samples, A and B, were microtomed Figure 1A shows the video image of microfilm A and the
and cut into films. The microfilms were placed onto glass location where the Raman line mapping was performed
slides and pentane was added to flatten the microfilms. (red line). The corresponding Raman line map is shown
The glass slide was then placed on the microscope stage in Figure 1B. The Raman line map is displayed in a 2-D
insert for Raman mapping. A Thermo Scientific DXR2 contour plot and reveals the compositional changes across
Raman Microscope was used for all analyses. A 532 nm different layers (y-direction). The rainbow color scheme
laser was used and the laser power was set at 5 mW. A of the contour plot reflects the intensity of the Raman
100× objective and a 25 μm confocal pinhole were used. peaks, with red indicating the highest intensity and blue
The Raman line map (27.4 μm) was collected with a 0.2-μm the lowest. Based on the Raman spectral changes in the
step size and contains 138 spectra. The Raman area vertical direction, seven distinct layers were identified by
map (3 μm [X] x 20 μm [Y]) was collected with 1 μm (X) library search: the layers 1, 3, 5 and 7 are polyethylene (PE),
and 0.2 μm (Y) step sizes and contains 584 spectra. Data the layers 2 and 6 are polypropylene (PP), and the layer 4
acquisition and processing were done by Thermo Scientific is polyvinyl alcohol (PVA). The representative Raman
OMNIC for Dispersive Raman software suite. spectra for those layers are shown in Figure 1C. No cross-
contamination between the adjacent layers was observed,
Results and discussion demonstrating an adequate spatial resolution to distinguish
Identification of the layers of polymer laminates using Raman adjacent layers. The thickness of the thinnest layer (Layer 4)
microscopy can be done by confocal depth profiling and was estimated to be 1.2 μm based on the full width at half
cross-section analysis. While depth profiling is advantageous maximum (FWHM) method (Figure 1D).3
in that it requires little to no sample preparation, cross-
section analysis via line/area mapping offers superior spatial
resolution, which lends itself to the identification of micron to
sub-micron layers in polymer laminates.

A B

C D

Figure 1. (A) Video image of microfilm A showing the region where Raman line mapping was performed (red line); (B)
Raman line map of microfilm A; (C) representative Raman spectra of each layer. PE = polyethylene, PP = polypropylene,
PVA = poly vinyl alcohol; and (D) Raman correlation profile obtained using a PVA reference spectrum.
Figure 2A-2C shows the 3-D Raman area images of correlation profiling using a PVA spectrum from the map
microfilm B. PE, PP and PVA spectra were carefully as the reference, and the result is shown in Figure 2D.
selected from the map as the references to construct The sharp peak at ~ 14.5 μm corresponds to the layer 4,
the Raman area images with correlation profiling. The with an estimated thickness of ~ 0.4 μm based on the
red color indicates a high correlation with the reference FWHM of the profile peak.3 It’s important to note that
material and the blue color indicates low correlation. the methodology described above relies heavily on the
Similar to microfilm A, seven layers are clearly identified: quality of the spectra as well as the spectral differences
the layers 1, 3, 5 and 7 are PE, the layers 2 and 6 are PP, between adjacent layers. The layer thickness derived from
and the layer 4 at the center is PVA. In order to estimate the correlation profile should therefore be considered an
the layer 4 thickness, a line map was first extracted estimate. Precise measurement of layer thickness requires
from the area map, delineated by the white, dashed such technique as scanning electron microscopy (SEM).
line in Figure 2C. The extracted line was then subject to

A B C

Figure 2. 3-D Raman correlation images for microfilm B. (A)-(C): Raman correlation images for PE, PP, and PVA, respectively. (D)
Raman correlation profile obtained using a PVA reference spectrum on a line extracted from the area map (white dashed line on C).

Diffraction-limited resolution in optical microscopy is often sample that is optically denser than air, the axial resolution
empirically assessed by the Rayleigh criterion shown is usually 4-6 times lower the lateral one. For a 532 nm
below, where d denotes the Rayleigh criterion, λ is the laser and a 100× objective (N.A. = 0.90), assuming the
laser wavelength, η is the refractive index of the immersion/ refraction index 1.5 for the laminates, the theoretical lateral
mounting media, and N.A. is the microscope objective spatial resolution and axial resolution are approximately
numerical aperture. 0.4 μm and 2 μm, respectively. However, many factors
such as scattering of the laser/Raman photons and
interaction with interfaces in the sample can reduce this
resolution. Another important consideration in instrument
configuration is the pinhole size. The confocal pinhole acts
as a spatial filter, by allowing the Raman spectrometer to
look into a smaller spatial domain than with a conventional
configuration without the pinhole, attenuating the out-
As indicated in the equations, both lateral and axial Rayleigh of-focus regions of the sample.2 In the current study, the
criteria are directly proportional to the wavelength, but combination of a 532 nm laser, a 100× objective with an
inversely proportional to the objective numerical aperture. N.A. of 0.9, and a 25 μm confocal pinhole enabled the
In addition, the axial Rayleigh criterion is proportional to the resolution of a 0.4 μm-thick PVA layer, approaching the
refractive index of the material under investigation. For any theoretical limit of the lateral spatial resolution.
Conclusions
In this communication, we have demonstrated the
analysis of polymer laminates using Raman line and area
mapping. In both cases, seven layers were identified.
With a properly configured Raman microscope, it was
possible to resolve a thin polymer layer with a thickness
(≈0.4 μm) that’s very close to the theoretical spatial
resolution limit.

References
1. Guillory P., Deschaines T. Henson P, Materials Today,
2009, 12(4), 38-39.

2. Rzhevskii A, Ibrahim M, Ramirez J, Macisaac C.,

Considerations and Techniques for Optimizing
Raman Spectral and Spatial Information, Thermo
Scientific White Paper 52699, 2015.

3. Guillory P., Deschaines T., Henson P. Confocal

Raman microscopy analysis of multilayer polymer
films, Thermo Scientific Application Note 51718,
2008.

Find out more at www.thermofisher.com/Raman

©2017 Thermo Fisher Scientific Inc. All trademarks are the property of Thermo Fisher Scientific and its
subsidiaries unless otherwise specified. AN52326_E 12/17M
 APPLICATION NOTE AN52818

FT-Raman mapping of multicomponent

solid dosage forms
Author Introduction
Ron Rubinovitz, Ph.D., Thermo FT-Raman spectroscopy is a
Fisher Scientific, USA complimentary technique to mid-IR
for material analysis. Compared to
Keywords FTIR, FT-Raman is compatible with
FT-Raman, microscopy, mapping, glass and requires minimal sample
PCA preparation. The use of a 1064 nm
laser in FT-Raman effectively alleviates
Thermo Fisher Scientific the fluorescence often observed when
solutions visible laser excitation is used. The
Thermo Scientific™ Nicolet™ iS™50 combination of FT-Raman spectroscopy
Figure 1: Nicolet iS50 FTIR spectrometer
FTIR, FT-Raman module, Thermo with mapping lends itself well for mixture shown with the iS50 Raman accessory in
Scientific™ OMNIC™, Atlμs™ and analysis by providing both chemical and the main sample compartment.
Specta™ software spatial information within a sample. However,
even a moderately sized map typically contains hundreds of spectra,
making the use of automated spectral analysis an important tool for the
identification of multiple components. The challenge is further exacerbated
by the disparity in component distribution within a mixture, especially when
spectral features only occur in small sections of a map and are obscured by
stronger matrix spectra. In those cases, a statistical tool such as principal
component analysis (PCA) is often required for optimal analysis.

In this note, a complete workflow to identify the multiple components

within an over-the-counter (OTC) drug tablet using FT-Raman mapping
is described. The capability of PCA to extract additional compositional
information from the measured spectra is also demonstrated.

Experimental
The cross section of an OTC bilayer tablet containing a blend of excipients and
an active pharmaceutical ingredient (API) was examined in this study. Spectra
were measured at 8 cm-1 resolution and 16 scans at each measurement point
using a Nicolet iS50 FTIR spectrometer, configured with a CaF2 beamsplitter,
InGaAs detector and iS50 Raman accessory (Figure 1). The laser wavelength
was 1064 nm with a spot size of approximately 50 microns and was applied
at a power of 500 mW. The area mapped was approximately 2 x 3 mm at 45
microns steps (Figure 2). Analysis of spectral maps was carried out using the
OMNIC Atlμs and OMNIC Specta software.
 Position (micrometers)= -143 µm; -15745 µm; Number: 185
Composite for Match 1

+
a-Lactose Monohydrate

Beta-Cyclodextrin Hydrate

Cellulose, Powder, Ca. 20 Micron

Figure 2: Cross section of tablet and area mapped.

Results and discussion

Preliminary Spectral Analysis
Figure 4: Multi-component search results for a spectrum from the
The area mapped consisted of three readily identifiable
Layer 2 Blend containing multiple excipients. The comparison of
regions. Two layers of the tablet could be visually the measured spectrum to the composite library match spectrum
distinguished while the third contained white particles is shown in the top of the figure, and the individual components are
shown below.
with an approximate dimension of 250 x 800 microns.
Representative spectra from each of these regions are Creating chemical maps
shown in Figure 3. Three chemical maps were generated for the API, the
Layer 2 Blend and mannitol, using 1755–1730 cm-1 (peak
Mannitol Region #1204
API Region #1191
area), 1334 cm-1 (peak height) and 875 cm-1 (peak height),
Layer 2 Blend #295 respectively. The results are shown in Figure 5. From
these maps based on spectral features unique to each
material, information on the component distribution and
interdependencies can be readily visualized.

Principal component analysis

As described above, an initial inspection of the sample
indicated three main spectral regions: API particles,
Figure 3: Representative map spectra from different regions of the a mannitol layer and the Layer 2 Blend. However, a
tablet. PCA analysis yielded four distinct components, as
Through a library search, the spectra of the first layer summarized in Figure 6. By visual inspection and library
yielded very high match values to mannitol. The distinct
comparison, the first three components modeled
mannitol, the Layer 2 Blend and the API, respectively.
spectra from the white particles were also unequivocally
The fourth component was an excellent library match
assigned to the active pharmaceutical ingredient (API) of
to acesulfame potassium, a sugar substitute known
the tablet. The spectra of the second layer are matched to to be used as an excipient in pharmaceutical dosage
lactose monohydrate in the library search, but with relative formulations. Closer evaluation of the map spectra
low match values. Multi-component library searches were confirmed the presence of this material in some very
subsequently performed on the spectra from this layer small regions of the map. It is also worthy to point out
(referred to henceforth as the Layer 2 Blend). The search that the PCA analysis was able to extract an almost pure
results (Figure 4) reveal that this layer of the tablet is a blend spectrum of acesulfame potassium, even though its
of excipients consisting chiefly of lactose monohydrate, beta features were heavily obscured by those from the Layer 2
cyclodextrin and microcrystalline cellulose. Blend spectrum in the experimentally measured spectra.
API Layer 2 Blend Mannitol
Figure 5: Maps created
based on peak heights
and area. High to low
intensities in the maps
are described by using
the colors red to blue.
Conclusions maps calculated from peak heights or areas are sufficient.
The analysis presented here demonstrates that However, as demonstrated in the case of acesulfame
recognition and spatial distribution of key materials in potassium, where a minor component occupies a very
pharmaceutical dosage forms can be accomplished small portion of the total map area and has physical
through FT-Raman mapping. The multi-component dimensions too small to permit the acquisition of a “pure”
search method is effective in deconvoluting the spectra to spectrum, PCA is a powerful analysis tool to detect and
yield the main components, when the measured spectra identify that component. The described workflow and
for each component have constant relative ratios, and its associated principles are applicable for comparing
the pure spectra of the components are available in the already-marketed drugs and providing valuable
library. When considering a spectral map containing information in pre-formulation development, formulations
known components with well isolated spectral features, assessment, scale-up and failure mode analysis.

Mannitol Region #1207 Layer 2 Blend #934

Principal Component 1 Principal Component 2 [Inverted]
D-Mannitol (Library)

API Region #1191 Acesulfame Region #1227

Principal Component 3 Principal Component 4
Acesulfame K (Library)

PC1 PC2 PC3 PC4

Mannitol Layer 2 Blend API Acesulfame Potassium

Figure 6: Top spectral plots show the spectra of the four identified components calculated by PCA: the experimentally
measured Raman spectra (blue), the PCA components (red), and library spectra (magenta) when available. Maps show
the distribution of each component across the measured area.

Find out more at www.thermofisher.com/ftir

©2018 Thermo Fisher Scientific Inc. All trademarks are the property of Thermo Fisher Scientific and its
subsidiaries unless otherwise specified. AN52818_E 04/18M
APPLICATION NOTE

No. AN52910
Characterizing amber and its imitations with
dispersive Raman spectroscopy
Author:
Alex Rzhevskii, Thermo Fisher Scientific,
Madison, WI

Key words

Raman spectroscopy, Raman microscope, dispersive

Raman, amber, copal, resin

Introduction
Amber is a tree resin that has been fossilized over
millions of years through polymerization, cross-linking
of the long carbon chain organics, and expulsion of
volatiles within the resin. Amber has a wide variety of
applications, ranging from its traditional use in jewelry,
art and craft objects, and folk medicine, to modern
applications in skin care products, perfumes, and While FT-Raman spectroscopy, in the near-infrared (1064
blood conservation devices. Authentic amber is a highly nm) region, has been used successfully to analyze amber1-3
demanded commodity and its high value has fostered an because it minimizes the substantial florescence originating
extensive imitation market that abounds with both natural from the amber resin, dispersive Raman spectroscopy
and synthetic resins. Copal is an example of one of the has important advantages over FT-Raman. The use of
natural resins that is used as an amber imitation. It is a lasers of shorter wavelength (e.g., 780 nm), especially,
less mature tree resin product that has not gone through in Raman microscopy allows for sample analysis on a
the entire fossilization process and, as a result, it is less sub-micrometer scale and results in significantly faster
expensive than amber. spectral acquisition, because of the high laser power
density developed in the small focused laser spots coupled
Distinguishing authentic amber from its imitations with more effective Raman scattering that is inversely
is important for scientific, commercial and personal proportional to the 4th power of the laser wavelength4.
reasons. There are several simple chemical and physical Furthermore, confocal Raman microscopy can provide
tests that can be used to authenticate amber. These useful information about the chemical composition of the
tests include solubility in organic solvents, floating in relict biomaterials and inorganic inclusions5.
salt water for testing specific gravity, assessing melting
points, and even evaluating hardness using scratch In this note, the spectra of amber and some of its
tests. The main limitation of those tests arises from common imitations obtained using a dispersive Raman
the possible compromise in specimen integrity, which spectrometer are compared and discussed.
can be prohibitive when dealing with items of artistic,
archeological or historical significance. To that end,
Raman spectroscopy offers an attractive solution to the
identification and characterization of amber, largely due to
its non-destructive sample handling.
Experimental Discrimination between amber and copal is slightly more
Samples of authentic and altered amber and its common involved because of their chemical similarities. Figure 2
imitations were used in this study, including: 1) Baltic shows the Raman spectra of amber and two copal samples
amber from a contemporary bead necklace, 2) a metal- of different origins. The most conspicuous features in
incrusted ear-ring made of, supposedly, a synthetic resin, all spectra are the bands at 1646 cm−1 and 1450 cm−1,
3) a pipe bit made of melted and pressed Baltic amber; resulting from the exomethylene stretching ν(C=C), and
4) West African copal, 5) Indonesian copal, 6) a sandarac the scissoring type deformation δ(CH2) vibration in resin
from Australia, 7) a pine tar, 8) a mastic resin. structure, respectively. It has been shown [2-4] that the
intensity ratio of the bands at 1646 cm−1 and 1450 cm−1
Raman spectra were acquired with a Thermo Scientific (/1646 ∕/1450) is a good indicator of the maturity of fossilized
DXR™ Raman microscope using 780 nm laser excitation resins. In general, the ratio is >1 for immature resins such
and 14 mW of power at sample. The typical exposure as copal, and <1 for mature resins such as amber. Our
time was set to 5 minutes to maximize signal-to-noise experiment results conform to that criterion. In Figure 2,
ratio. The real-time autofluorescence correction was the peak height ratio /1646 ∕/1450 is about 0.8 for amber (1) and
enabled to minimize fluorescence effects in the spectra. 2.5 and 1.1. for copal (2) and (3), respectively. In addition,
No further smoothing of spectral data was applied. there are noticeable differences between amber and copal
in the spectral region below 1400 cm-1, which is dominated
Results and discussion by a series of less intensive peaks, especially, in the 800-
Distinguishing amber from its synthetic forgeries by 500 cm-1 range. It was found that thermal and mechanical
Raman spectroscopy is rather straightforward. Figure 1 alteration of the amber leads to noticeable spectral changes
shows the “fingerprint” region of the Raman spectra of in the region of 1800 - 1400 cm-1, as shown in Figure 3.
an authentic Baltic amber and a synthetic resin, as well
as a reference spectrum of bisphenol-A. The spectrum
of the synthetic resin (Figure 1.2) bears little similarity
with that of amber (Figure 1.1), but has features that can
be unambiguously attributed to the bisphenol-A, the
component of an epoxy resin. Thus, in the “fingerprint”
spectral region, amber can be readily distinguished
from synthetic plastics due to a significant difference
in chemical composition and the identification of the
synthetic plastics can be conducted using a search
against extensive Raman spectral libraries of polymers.

Figure 2: Raman spectra of (1) melted and pressed Baltic amber from a pipe bit; (2)
Indonesian copal; and (3) West African copal. The spectra are normalized to the most
intense spectral bands.

Figure 1: Raman spectra of: (1) authentic Baltic amber from a contemporary bead
necklace; (2) synthetic resin from an ear-ring; and (3) bisphenol-A from the Nicolet
Standard Collection of Raman Spectra. The spectra are normalized to the most intense
spectral bands.

Figure 3: The Raman spectra of Baltic amber from bead necklace (red) and melted and
pressed amber from a pipe bit (brown).
The band at 1646 cm-1 becomes broader, and the intensity The spectral characteristics of amber are demonstrated to
of the shoulder at about 1615 cm-1 associated with the be vastly different from, for example, epoxy resin often used
stretching of aromatic C=C increases in the spectrum of for amber imitation. The fluorescence typical for amber,
the processed amber. In [6], these changes have been copal and other natural resins are effectively suppressed
observed and attributed to the increase of the aromaticity by using a 780-nm excitation, or it can be removed with
as the result of the thermal impact to the material. a real-time automated fluorescence correction procedure
available in Thermo Scientific™ OMNIC™ software. The
Raman spectroscopy can be used to differentiate amber use of dispersive Raman instrumentation along with visible
from other types of natural resins that look similar to amber. lasers enabled shorter sampling times in comparison with
Figure 4 shows the Raman spectra of three different resins. FT-Raman, where spectral acquisition times as long as
In the spectrum of mastic resin (1), the ratio /1646 ∕/1450 is <1 hours are often required5.
indicating its similarity to amber. However, the band at 1646
cm-1 in the spectrum of mastic resin exhibits a prominent References
shoulder at 1710 cm-1 less noticeable for amber (Figure 2.1)
that is ascribed to the stretching mode ν(C=O). For the pine 1. Beck C.W. Spectroscopic investigation of amber,
tar (2) and sandarac from Australia (3), the ratio /1646 ∕/1450 Applied Spectroscopy Reviews, Vol. 22, 57–110, 1986
is >1 and ≈1, respectively, typical for natural resins. In the
spectrum of pine tar (2), there is also a salient side peak at 2. Edwards H.G.M., Farwell D.W. Fourier transform-
about 1615 cm-1. Raman spectroscopy of amber, Spectrochimica Acta,
Part A, Vol. 52, 1119-1125, 1996.

3. Brody R.H., Edwards H. G.M , Pollard A.M., A

study of amber and copal samples using FT-Raman
spectroscopy, Spectrochimica Acta, Part A, Vol. 57,
1325–1338, 2001.

4. Vandenabeele P., Grimaldi D.M., Edwards H.G.M.,

Moens L. Raman spectroscopy of different types of
Mexican copal resins, Spectrochimica Acta, Part A,
Vol. 59 2221-2229, 2003.

5. Edwards H.G.M., Farwell D.W., Jorge Villar S.E. Raman

Figure 4: The Raman spectra of: (1) a mastic resin; (2) a pine tar; and (3) a sandarac microspectroscopic studies of amber resins with
from Australia. The spectra are normalized to the most intense spectral bands. insect inclusions, Spectrochimica Acta, Part A, Vol. 68,
1089–1095, 2007.
All the collected spectra feature high signal-to-noise ratios
and sufficient spectral resolution, allowing for the observation 6. Badea G.I., Caggiani M.C., Colomban P., Mangone A.,
of fine spectral details for differentiation purpose. Teodor E.D., Teodor E.S., Radu G.L. Fourier Transform
Raman and Statistical Analysis of Thermally Altered
Conclusions Samples of Amber, Applied Spectroscopy, Vol. 69,
Dispersive Raman spectroscopy is a practical tool for non- 1457-1463, 2015.
destructive identification and authentication of amber and
its imitations made of natural and synthetic resins.

Find out more at www.thermofisher.com/raman

For Research Use Only. Not for use in diagnostic procedures. © 2016 Thermo Fisher Scientific Inc. All rights
reserved. All trademarks are the property of Thermo Fisher Scientific and its subsidiaries unless otherwise specified.
AN52910_E 12/16M
 APPLICATION NOTE AN52984

Confocal Raman analysis of a

transdermal nicotine patch by a DXR2 Raman Microscope
Author Introduction
Mohammed Ibrahim, Ph.D. Transdermal patches are medication delivery systems that adhere to the skin
Thermo Fisher Scientific, USA and deliver drugs through a controlled release into the skin and then into the
bloodstream. Nicotine transdermal patches are a form of nicotine replacement
Application benefit therapy used to aid in smoking cessation.
Confocal Raman microscopy
provides a facile means to analyze Transdermal patches have multiple layers: the release liner, the adhesive, the
multilayer polymers and to identify membrane, and the backing. The release liner protects the adhesive and drug
the active pharmaceutical ingredient formulation and is removed prior to application. The adhesive adheres the patch
(API) in a nondestructive manner to the skin. The membrane layer (or layers) contains the drug and controls the
with minimal sample preparation. rate of diffusion of the drug out of the patch and into the skin. The backing is the
outermost layer of the patch and protects the formulation while the patch is worn.
Keywords
Confocal Raman microscopy, It is important to understand the composition of a transdermal patch to perform
Transdermal patch, Chemical quality control and product improvement, and to monitor the drug release
mapping, Multilayer polymer mechanism. Traditional methods of analysis are typically destructive, either
requiring dissolving the sample for extraction (e.g., GC-MS, LC-MS) or physically
cross-sectioning the patch (e.g., SEM, TEM). Using confocal Raman analysis by
the Thermo Scientific™ DXR™2 Raman Microscope, the layers of a commercially
available transdermal nicotine patch were examined without sample destruction.

Experimental
A transdermal nicotine patch sample was mounted onto a gold-coated
microscope slide with the backing layer facing the microscope objective, and
the release liner at the bottom. Raman confocal line depth profiling and area
depth profiling were performed with a DXR2 Raman microscope (Figure 1)
using 532 nm laser, 5 mW laser power at the sample, 50X objective, 25 µm
Figure 1: DXR2 Raman microscope
confocal pinhole aperture, and with auto exposure (S/N = 200). For the line
depth profiling (Z-profiling), a depth of 220 μm was probed by using a 5 μm
step size (containing 45 points or spectra). For the area depth profiling, a
vertical X-Z area of 120 μm (X) by 245 μm (Z) was probed by using a 5 μm step
size along the Z-axis and 20 μm along the X-axis. The area is composed of
350 points or spectra.
To correlate an optical image with the Raman profile, a at half maximum (FWHM) value method described in
small cross sectional piece of the nicotine patch was the Thermo Scientific Application Note 517181. Figure 2b
mounted vertically on the glass slide. shows the optical image of the physically cross-sectioned
sample. The optical image provides a visual representation
The Thermo Scientific™ OMNIC™ for Dispersive Raman of the layer thickness based on contrast and although it is
Software Suite, which includes Thermo Scientific™ comparable to the chemical image, some of the layers are
OMNIC™ Atlμs™ Software, was used for data collection, not easily identifiable.
processing, and analysis, including identification of the
layer composition and estimation of the layer thickness. Figure 3 shows the representative Raman spectra for
The multi-component spectral analysis was performed each layer in the Raman line depth profile. The chemical
using Thermo Scientific OMNIC Specta Software. composition of each layer was initially identified
by performing a library search against the Thermo
Results and discussion Scientific High-Resolution Raman Polymer Library using
A Raman line depth profile map obtained using the OMNIC OMNIC software.
software suite is shown in Figure 2a. In the figure, the depth
profile is displayed in a 2-D contour map and reveals the Layer 1, the backing is poly(ethylene terephthalate) or
changes in the Raman peaks (vertical lines) as the focal PET. Layers 2 and 4 are composed of microporous
point probes through the layers. The rainbow color scheme polyethylene (PE) that acts as a controlled release agent
of the contour map shows the intensity of the Raman peaks for nicotine, which is located at the ethylene/vinyl acetate
as color, with red indicating the highest intensity and blue copolymer (EVA) reservoir layer (layer 3). Layer 5 is a
the lowest intensity. A total of six layers were identified polyisobutylene (PIB) adhesive and PET. Layer 6, the
with thicknesses ranging from ~15 μm to ~75 μm. The release liner is also composed of PET.
thickness of each layer was estimated by the full width

Figure 2: a) Raman line depth map (Z-map)

for the transdermal nicotine patch. The
dashed white lines indicate the approximate
depth of each of the six layers. b) An optical
image of the cross section.

Figure 3: Representative Raman spectra for

each layer of the patch.
The presence of nicotine in the EVA reservoir layer Figure 5 shows correlation profile images obtained from
(layer 3) was identified by the OMNIC Specta software the Raman confocal area depth (X-Z) profiling of the same
using the Multi-Component Search option, as shown in nicotine patch sample. Each of the images was obtained
Figure 4. The patented Multi-Component Search allows by using the correlation between the map spectra and
one to identify constituents in mixture samples without the reference Raman spectrum of each layer. The red
the tedious manipulations required by the traditional color in the images indicates the highest correlation to a
search and subtract processes. Not only did the search specific reference spectrum, whereas the blue indicates
identify S-(-)-nicotine as the second component (green in no correlation. For example, the EVA+nicotine correlation
Figure 4), in addition to the EVA (orange), it also provided image (Figure 5, right) shows a single red colored layer
the composite value (%) of the two components. The indicating that this layer has the highest correlation to the
composite value is a scaling factor indicating the amount EVA+nicotine reference spectrum, thus showing that this
of each found reference represented in the composite is the layer where these two components are located. The
spectrum. It should be noted, however, that the composite presence of 6 layers in these images is consistent with
values cannot be taken as a quantitative percentage of results of the confocal line depth profile analysis (Figure
the materials, which requires measurements by a primary 2). In addition, the X-Z correlation images here reveal
quantitative analytical technique. relatively uniform thickness of the layers.

Figure 4: OMNIC Specta Multi-Component Search identified the

presence of (S)-(-)-nicotine (~36%) in addition to the EVA co-polymer
in layer 3.

PET PIB PE EVA + nicotine

Figure 5: X-Z correlation profile images generated by the reference Raman spectrum of each layer.
Conclusion References
Existing techniques for analyzing a multilayer polymer like 1. Guillory P., Deschaines T., Henson P. Confocal Raman
a transdermal nicotine patch are labor intensive, time- microscopy analysis of multilayer polymer films,
consuming, and would require destruction of the sample. Thermo Scientific Application Note 51718, 2008.
Using confocal Raman analysis, the analysis was quick,
and the sample integrity was preserved. The chemical
composition of each layer was determined using OMNIC
for Dispersive Raman and OMNIC Specta software suites.
OMNIC Specta Multi-Component Library Search identified
the constituents in layer 3, a mixture of nicotine and an EVA
co-polymer, which is easier than tedious manipulations
required by traditional search and subtract processing.

Find out more at thermofisher.com/Raman

©2017 Thermo Fisher Scientific Inc. All rights reserved. All trademarks are the property of Thermo Fisher
Scientific and its subsidiaries unless otherwise specified. AN52984_E 10/17M
 APPLICATION NOTE AN53001

In situ density determination of polyethylene in

multilayer polymer films using Raman microscopy
Authors Introduction
Mohammed Ibrahim, Ph.D. Polyethylene (PE) is one of
Herman He, Ph.D. and the most common plastics
Rui Chen, Ph.D. in the world with an annual
Thermo Fisher Scientific global production of around
Madison, WI, USA 80 million tons.1 Unlike other
commercial polymers, PE can be
Keywords manufactured across a range of
Polyethylene (PE), density, densities that are related to the
Raman Microscopy, DXR2 spacing between the polymer chains.2 For examples, Thermo Scientific DXR2
Raman Microscope, Partial High-Density PE (HDPE, 0.941-0.965 g/cm3) is Raman Microscope

Least Squares (PLS) mainly comprised of linear PE chains that can

closely approach each other, thus creating very densely packed networks.
Application benefits Conversely, Low-Density PE (LDPE, 0.910-0.940 g/cm3) has excessive
Polyethylene density determination branching that causes a less compact molecular structure. Linear Low-
based on Raman microscopy Density PE (LLDPE, <0.940 g/cm3), too, has a large number of branches, but
and Partial Least Squares (PLS) the branches are shorter than those in LDPE. Because the density of PE can
regression is applicable for both be controlled and varied, it provides an important mechanism to control its
pellet and film samples. The properties to suit different applications. PE density is therefore a vital part of
confocal capability of Raman the material characterization and selection process.
microscopy allows for in situ density
determination of PE layers within There are several standard methods for PE density measurement, such
multilayer polymer films, without the as ISO 1183-1/ASTM D792 (immersion method)3, ISO 1183-2/ASTM D1505
need for tedious and challenging (density gradient method),4 and ASTM D4883 (ultrasound method)5. These
sample preparations. methods, however, are primarily geared towards PE samples in a “pure”
form such as pellets and single-layer films. Challenges arise when PE is
Thermo Fisher Scientific present in multilayer films, which are widely used in food, pharmaceutical
solutions and consumer product packaging. Extensive sample preparations,
• DXR2 Raman Microscope including microtoming and separation of layers using solvents, are required
to isolate the PE layer(s) before analysis, which can be labor-intensive and
• OMNIC software
time-consuming.6
• TQ Analyst software
To that end, Raman microscopy could offer an in situ was then used for final analysis. A 532 nm laser was
density determination of PE layers in multilayer films by used with 2 mW laser power at the sample. A 10x
leveraging its confocal capability. Raman spectroscopy objective and a 50 μm slit aperture were used to obtain
has long been utilized to investigate the structure of more representative spectra from the samples. Total
polymers, including molecular conformation, orientation, acquisition time for each spectrum was 30 seconds (3
and crystal structure.7-9 Combined with multivariate second exposure x 10 exposures). For the transdermal
analyses such as principal component analysis (PCA) nicotine patch sample, Raman confocal line depth
and partial least squares (PLS) regression whereby profiling was performed using a 532 nm laser, 5 mW
statistically observed spectral variations are correlated laser power at the sample, 50x objective, 25 μm confocal
with known sample properties, crystallinity, density pinhole aperture, and with auto exposure (S/N = 200). A
and melting point can be predicted.7 Existing literature, depth of 220 μm was probed by using a 5 μm step size
however, has also been limited to bulk PE pellets. In (containing 45 points or spectra).
our previous report,10 we expanded the scope of the
methodology to include PE films. It is demonstrated that Thermo Scientific™ OMNIC™ software was
Raman microscopy in combination with a PCA based used for instrument control and data acquisition.
discriminant analysis can qualitatively distinguish HDPE Thermo Scientific™ TQ Analyst™ software was used for
and LDPE in both pellet and film forms. In this application chemometric analysis of the Raman data.
note, we will describe the development of a Partial Least
Squares (PLS) model to quantitatively determine the Results and discussion
density of PE layers in both single-layer and multilayer Raman spectra of HDPE, LDPE, and LLDPE
films using Raman microscopy. Representative Raman spectra of HDPE, LDPE, and
LLDPE samples are shown in Figure 1. There are
Experimental noticeable differences among three types of PE samples
A total of 25 PE samples (12 pellets and 13 films) in both CH stretching (2900-3100 cm-1) and CH2 bending
with different known densities (Table 1) were used for and twisting (1250-1500 cm-1) regions. The intensity of
developing a PLS model for density determination. All the symmetric CH2 stretching mode at 2848 cm-1 (relative
samples were used as received. A clear transdermal to the asymmetric CH2 stretching mode at 2882 cm-1)
nicotine patch sample was used for identifying the PE decreases in the order of LLDPE > LDPE > HDPE
layers and in situ determination of PE layer densities. The (Figure 1A). In the CH2 bending and the CH2 twisting region
sample was mounted onto a gold-coated microscope (Figure 1B), the intensity of the CH2 bending mode at
slide with the backing layer facing the microscope 1416 cm-1 (relative to the CH2 bending mode at 1440 cm-1)
objective, and the release liner at the bottom. is higher for HDPE than for LDPE. The 1416 cm-1 peak is
completely absent for the LLDPE film sample (#19). This
A Thermo Scientific™ DXR2™ Raman Microscope was observation agrees with the previous reports that the
used for collecting Raman data. For each of the pellet 1416 cm-1 and 1440 cm-1 peaks are indicators of crystalline
samples, Raman spectra were collected from 3 different and amorphous PE phases, respectively.7,8 The higher the
pellets and averaged. For each of the film samples, crystallinity, the higher the density. Since the CH2 bending
Raman spectra were collected from 3-4 locations across region (1400-1500 cm-1) is sensitive to the density of PE, it
the surface of the sample. An averaged spectrum was selected for subsequent quantitative analysis.

Figure 1: Representative Raman spectra

of HDPE, LDPE, and LLDPE samples.
(A) Full spectral range in stack view. (B)
CH2 bending and CH2 twisting region in
overlay view.

A B
 Data processing through statistical analysis. It uses spectral covariance and
Peak area for pathlength correction option in the TQ factorial analysis to extract significant and relevant chemical
Analyst software was used to normalize Raman spectral information from sample spectra as factors, then correlate
intensities. The peak area of the CH2 bending mode at them with sample properties such as concentration,
1440 cm-1 was used for the normalization (1422-1452 cm-1 crystallinity and density. A total of 20 samples, a mix of
range, Figure 2A). An averaged two-point baseline pellets and films, were used as the calibration standards.
correction was used to account for baseline shifts/noise. Five additional pellet and film samples with density
values spread across the density range of the samples
Developing a PLS model for PE density determination were selected as the validation standards (Table 1). A
Partial Least Squares (PLS) algorithm11 from the TQ Analyst spectral range of 1400-1500 cm-1 with averaged two-point
software was used to develop a model for PE density correction for baseline (Figure 2B) was used in the method.
determination. PLS is a quantitative regression algorithm

Figure 2: (A) Peak Area Ratio to A B

normalize Raman spectral intensities,
showing the CH2 bending peak region for
normalization. (B) Spectral range used
for PLS calibration. In both cases, an
averaged two-point option was used for
baseline correction.

Actual Density Calculated Density Difference

Sample # PE Type Usage % Difference
(g/cm3) (g/cm3) (g/cm3)
1 0.9460 HDPE Calibration 0.9454 0.0006 0.06%
2 0.9505 HDPE Calibration 0.9486 0.0019 0.20%
3 0.9510 HDPE Calibration 0.9509 0.0001 0.01%
4 0.9470 HDPE Validation 0.9523 -0.0053 -0.56%
5 0.9620 HDPE Calibration 0.9564 0.0056 0.58%
Pellets

6 0.9600 HDPE Calibration 0.9594 0.0006 0.06%

7 0.9195 LDPE Calibration 0.9236 -0.0041 -0.45%
8 0.9170 LDPE Calibration 0.9183 -0.0013 -0.14%
9 0.9235 LDPE Calibration 0.9250 -0.0015 -0.16%
10 0.9300 LDPE Calibration 0.9264 0.0036 0.39%
11 0.9235 LLDPE Calibration 0.9295 -0.0060 -0.65%
12 0.9185 LLDPE Calibration 0.9259 -0.0074 -0.81%
13 0.9496 HDPE Calibration 0.9568 -0.0072 -0.76%
14 0.9606 HDPE Calibration 0.9556 0.0050 0.52%
15 0.9460 HDPE Calibration 0.9456 0.0004 0.04%
16 0.9247 LDPE Calibration 0.9225 0.0022 0.24%
17 0.9258 LDPE Calibration 0.9241 0.0017 0.18%
18 0.9297 LDPE Calibration 0.9276 0.0021 0.23%
Films

19 0.8598 LLDPE Calibration 0.8604 -0.0006 -0.07%

20 0.8650 LLDPE Calibration 0.8623 0.0027 0.31%
21 0.8881 LLDPE Validation 0.8812 0.0069 0.78%
22 0.9008 LLDPE Calibration 0.8993 0.0015 0.17%
23 0.9040 LLDPE Validation 0.9072 -0.0032 -0.35%
24 0.9236 LLDPE Validation 0.9215 0.0021 0.23%
25 0.9367 LLDPE Validation 0.9349 0.0018 0.19%

Table 1: PE samples and their densities. *Rows highlighted in green are the samples used for validation.
Figure 3 shows the calibration results for PE densities important to note that the percent differences do not
obtained with the 3-factor PLS model. The inset is the exhibit any bias between pellets and films, indicating that
Predicted Residual Error Sum of Squares (PRESS) plot. the sample form (pellets vs. films) has no bearing on the
In the current case, a 3-factor model suffices as the model performance.
contribution from the 4th and 5th factors are negligible.
The calibration curve has a correlation coefficient of Density Determination of PE films
0.9914. The RMSE (Root Mean Square Error) values are Figure 4 demonstrates an example of applying the
0.00360 for the calibration samples and 0.00432 for the Raman spectrum and the PLS model to predict the
validation samples, respectively. The results are also density of a PE film sample. The predicted density is
summarized in Table 1. For all 25 samples, the calculated 0.9014 g/cm3, showing a good agreement with the actual
densities are within ±0.81% of the actual values. It is density of 0.9008 g/cm3.

Figure 3: Calibration results for PEs of different densities using a PLS quantitative analysis. The  denotes calibration standards
and the + denotes validation standards. Inset is the PRESS plot for the PLS calibration. Three factors were used in the current PLS
calibration model.

Figure 4: Prediction of the density of a PE film sample using its Raman spectrum and the PLS model. The known density of the film is
0.9008 g/cm3 and the predicted density is 0.9014 g/cm3.
Figure 5A shows the Raman confocal depth profile of Conclusions
a clear transdermal nicotine patch. A total of 6 polymer Raman microscopy is a powerful analytical tool for PE
layers were identified, including two PE layers – PE density determination. Since PE chains in crystalline
layer 1 (part of the backing layer) and PE layer 2 (closer and amorphous domains exhibit unique Raman features
to the release liner). There are perceptible differences in in the CH2 bending region, a PLS model based on
the Raman spectra between the two PE layers (Figure 5B the Raman features in the 1400-1500 cm-1 region was
and 5C). Applying the PLS model, the densities are successfully developed. The model is applicable for both
determined to be 0.9150 g/cm3 for PE Layer 1 and pellet and film samples, showing a good agreement
0.9583 g/cm3 for PE Layer 2, placing PE layer 1 in the between actual and predicted density values. Applying
LDPE /LLDPE class and the PE Layer 2 in the HDPE the model to a real-world multilayer film containing two
class. The classification of the PE layers based on the PE layers, the predicted density values correctly place
predicted densities conforms to other reports: LDPE/ the two layers into their respective PE classes. More
LLDPE is used in the occlusive backing layer for its importantly, the confocal capability of Raman microscopy
flexibility whereas HDPE is used as the rate-controlling allows for in situ density determination of PE layers within
membrane as an integral part of the reservoir diffusion multilayer polymer films, without the need for tedious and
control machanism.12-13 While the exact densities of the challenging sample preparations required by many other
two PE layers are not available, the results presented techniques. The presented methodology should be of
here nonetheless demonstrates the advantage of using interest for PE manufacturers as well those who perform
Raman microscopy combined with the PLS method for failure analysis, reverse engineering, and polymer
density determination. The confocal capability of Raman composites development.
microscopy allows for in situ PE density determination in
multi-layer films without the need to isolate the individual Acknowledgement
PE layers. The authors would like to thank Wanda Weatherford at
Chevron Phillips Chemical Company for providing PE
pellet samples and Rajesh Paradkar at Dow Chemical
Company for providing PE film samples.

A B

Figure 5: (A) Confocal Raman line depth map of a multilayer polymer patch, showing the presence of two types of PE layers. (B) Raman
spectra of the two PE layers in full spectral range. (C) Raman spectra of the two PE layers in the CH2 bending and CH2 twisting region.
The insets in (C) show the calculated densities of the two PE layers using the 3-factor PLS model. PET = poly(ethylene terephthalate),
EVA = ethylene/vinyl acetate copolymer, PIB = polyisobutylene.
References
1. Piringer O.G. and Baner A.L., ed. Plastic Packaging: Interactions with Food and Pharmaceuticals. 2nd ed.
Weinheim: Wiley-VCH; 2008.
2. Polyethylene, The Essential Chemical Industry – online, http://www.essentialchemicalindustry.org/polymers/
polyethene.html, retrieved on 11/27/2017.
3. (a) ISO 1183-1:2012, Plastics – Methods for determining the density of non-cellular plastics − Part 1: Immersion
method, liquid pyknometer method and titration method; ANSI New York, NY. (b) ASTM D792-13, Standard test
methods for density and specific gravity (relative density) of plastics by displacement; ASTM International, West
Conshohocken, PA, 2013.
4. (a) ISO 1183-2:2004, Plastics – Methods for determining the density of non-cellular plastics – Part 2: Density
gradient column method. ANSI New York, NY. (b)ASTM D1505-10, Standard test method for density of plastics by
the density-gradient technique; ASTM International, West Conshohocken, PA, 2010.
5. ASTM D4883-08, Standard test method for density of polyethylene by the ultrasound technique; ASTM
International, West Conshohocken, PA, 2008.
6. Mieth A., Hoekstra E., and Simoneau C. Guidance for the identification of polymers in multilayer films used in food
contact materials: User guide of selected practices to determine the nature of layers, EUR 27816 EN, 2016; doi:
10.2788/10593.
7. Sato H., Shimoyama M., Kamiya T., Amari T., Sasic S., Ninomiya T., Siesler H.W., and Ozaki Y. Raman spectra of
high-density, low-density, and linear low-density polyethylene pellets and prediction of their physical properties by
multivariate data analysis, J. Appl. Polym. Sci., 2002, 86, 443–448.
8. Strobl G.R. and Hagedorn W. Raman spectroscopic method for determining the crystallinity of polyethylene, J.
Polym. Sci. B Polym. Phys., 1978, 16, 1181-1193.
9. Williams K.P.J. and Everall, N. J. Use of micro Raman spectroscopy for the quantitative determination of
polyethylene density using partial least-squares calibration, J. Raman Spectrosc. 1995, 26, 427-433.
10. Ibrahim M. and He H. Classification of polyethylene by Raman spectroscopy, Thermo Scientific Application Note
AN52301, 2017.
11. Thermo Scientific Product Overview. TQ Analyst Software Chemometric Algorithms, 2009.
12. Kearney C. J. and Mooney D.J. Macroscale delivery systems for molecular and cellular payloads, Nat. Mater.,
2013, 12, 1004-1017
13. Allen L. V. and Ansel H. C., ed. Ansel’s Pharmaceutical Dosage Forms and Drug Delivery Systems, 10th ed.
Lippincott Williams & Wilkins; 2013.

Find out more at www.thermofisher.com/Raman

©2018 Thermo Fisher Scientific Inc. All trademarks are the property of Thermo Fisher Scientific and its
subsidiaries unless otherwise specified. AN53001_E 03/18M
APPLICATION NOTE

No. AN52959
Simultaneous rheology and Raman
spectroscopy during the melting and
recrystallization of polypropylene
Authors: Nathan C. Crawford and Figure 1:
David Drapcho The Thermo Scientific HAAKE
MARS XR RheoRaman system.

Introduction
Rheology is the study of the flow and deformation of
matter, including fluids to solid-like materials (and anything
in between). Rheological measurements are commonly
used to examine or induce bulk physical material
changes, such as melting, crystallization, gelation, and/
or polymerization etc. Raman spectroscopy, on the other
hand, is a vibrational spectroscopic technique that can
provide insights to these processes at the molecular
level, both chemically and morphologically. However,
until very recently1, Raman studies are often conducted In this note, a MARSXR RheoRaman system was used
ex-situ, whereby Raman spectra are acquired prior to to investigate the temperature-dependent melting and
and after the observed physical transformation. The true crystallization of polypropylene, as well as the isothermal
chemical/morphological changes driving these processes crystallization process. The melt and crystalline phase
are largely left uncaptured, leaving much room for data transitions of polymeric materials are commonly correlated
interpretation and speculation surrounding these dynamic with variations in viscous and elastic behavior during
physicochemical relationships. rheological analysis. In addition, these phase transitions
are often associated with spectral pattern changes
The Thermo Scientific™ HAAKE™ MARSXR (Figure 1) in characteristic Raman peaks during spectroscopic
is a fully integrated RheoRaman system that enables investigation. Measurements performed ex-situ are
simultaneous rheology and Raman spectroscopy often challenging to compare due to discrepancies
measurements. The seamless hyphenation of these two in temperature control, slight deviations in sample
techniques allows for real-time, in-situ measurements of composition, and differences in processing history. The
both physical and chemical/morphological properties with in-situ RheoRaman system, on the other hand, completely
great ease. This multimodal analytical tool offers several eliminates these discrepancies, allowing for a more valid
advantages compared to the traditional ex-situ approach. analysis of the melt and crystalline phase transitions from
First, data collection efficiency is greatly improved by both the macroscopic and molecular levels.
combining multiple one dimensional experiments into a single
multifaceted experiment. Secondly, because both techniques
are employed simultaneously, sample fidelity is preserved
and the transformation of the material is captured in real-
time. And finally, sample consumption is reduced, which
can be beneficial for new material/formulation development
where sample quantities are limited and/or expensive.
Materials and methods The iXR Raman spectrometer is free-space coupled to
Materials the rheometer with an optical train which uses two plane
Polypropylene (Ineos Olefins and Polymers, USA, R12C-00 mirrors to direct the incident laser into the RheoScope
random copolymer) pellets were used for this study. The module. Within the RheoScope, a series of mirrors directs
pellets were melted at 190 ºC on the rheometer to form a the laser beam into a 20x objective, where the laser light
continuous, disk-shaped specimen for testing. is focused through a 2 mm thick fused silica window
into the sample (perpendicularly to the flow or vorticity
Rheometer plane). Raman scattered light is collected in a 180º
Rheological measurements were performed using the backscatter geometry using the 20x objective, and back
Thermo Scientific™ HAAKE™ MARS™ III rheometer, into the spectrometer through the same optical train as
equipped with a 20 mm diameter stainless steel parallel- the incident laser (eventually to the spectrograph inside
plate rotor. In order to extract viscoelastic behavior of the spectrometer). Free-space coupling of the laser to the
the polypropylene, all measurements were conducted rheometer, and the iXR spectrometer design, allow easy
in the oscillatory mode. Oscillatory measurements were Raman excitation laser wavelength interchange to permit
performed at 1 Hz with a constant strain of 0.1%, while optimization of the laser wavelength to the sample (785 and
data were collected every 5 s. To soften the polypropylene 455 nm laser sets are also available).
and help it conform to the measuring geometry, all
samples were initially loaded at 190 ºC. For the melt-to- The sample is positioned between the silica window and the
crystallization phase transition study, the temperature was rotor geometry attached to the rheometer measuring head
decreased from 190 ºC to 30 ºC, at a rate of 5 ºC/min. For (Figure 2). The objective can be adjusted for interrogation
the isothermal crystallization studies at 138 ºC and 150 at different penetration depths within the sample, as well
ºC, temperature was rapidly decreased from 190 ºC to 10 as positioned at various radial locations along the optical
ºC above the target temperature. The temperature was slit (from the true center to outer edge of the sample). An
then slowly decreased until it reached the test temperature electrical heating element is positioned below the fused
(either 138 ºC or 150 ºC). The temperature was then silica window to provide temperature control during testing.
held constant for a maximum of 1 h (3600 s), while the Cooling was provided from a temperature-controlled
isothermal recrystallization process was observed. circulator with a 50:50 mixture of ethylene glycol and
water. All instrumentation is controlled through the Thermo
Spectrometer Scientific™ OMNIC™ and RheoWin software packages.
Raman spectroscopy measurements were
performed using the Thermo Scientific™
iXR™ Raman Spectrometer. The iXR
system employed a 532 nm, 10 mW
laser, a triplet spectrograph providing
Raman spectra over the range 3500 to
50 cm-1 Raman shift (Stokes) at 5 cm-1
resolution, and a CCD camera cooled
to -50 ºC. Alignment of the laser, Raman
scatter, and aperture selection within
the spectrometer were all software
controlled. The minimum exposure
collection time is 0.1 s. For the data
presented here, the exposure collection
time was 4 s and 2 sample exposures
were averaged per spectra collection.

Figure 2: Schematic diagram of the MARS XR RheoRaman system (showing side and
RheoRaman coupling
top views of the rheometer sample stage). The iXR Raman spectrometer is free-space
The Raman spectrometer (Thermo Scientific iXR Raman coupled to the MARS rheometer using plane mirrors that direct light into a 20x long-
Spectrometer) and rheometer (Thermo Scientific HAAKE working-distance objective. The objective focuses the incoming laser (green dashed
line) and collects the back-scattered Raman light (yellow) coming into and out of the
MARS III) are coupled together using the Thermo Scientific™
sample (which sits atop the rheometer stage). The optical path from the spectrometer
HAAKE™ RheoScope module (Figure 2). to the rheometer is enclosed in lens tubes (black). The optical slot in the bottom
heating plate permits passage of the laser beam and scattered Raman light.
The electric heater allows for a maximum temperature melt-to-recrystallization process2, while the spectral features
of 300 ºC and a minimum temperature of -5 ºC, with a in the 3050-2800 cm-1 range were integrated into a total peak
maximum heating/cooling rate of 10 ºC/min. An active area. These indicators (normalized helical chain vibration
electrical hood was also used to provide temperature peak height and overall CH stretching peak area) were then
control from above, eliminating the potential for a tracked throughout the melt- to crystalline-phase transition
temperature gradient within the sample (Figure 2). and overlaid with the in-situ rheology data (Figure 4).
The melt and recrystallization process of polypropylene
Results and discussion was probed rheologically using small amplitude oscillatory
The physicochemical and morphological relationships shear measurements (Figure 4), where the storage modulus
during polymer crystallization are of critical importance (G’) and loss modulus (G”) were measured as a function of
for polymer processing. Here we use the new MARSXR temperature. G’ and G” are measures of a material’s elastic
RheoRaman system to investigate the crystallization process and viscous behavior, respectively. A liquid-like material will
of polypropylene (PP). Representative Raman spectra for be more viscous than elastic (i.e., viscously dominated),
PP at three different temperatures are shown in Figure 3. In and as a result, G” will be greater than G’. Conversely, a
general, the spectrum at 53 ºC shows sharp peaks across solid-like material will display more elastic than viscous
the examined spectral range. These sharp spectral features behavior (i.e., elastically dominated), where G’ will be
suggest a high degree of conformational order, indicative of greater than G”.
semicrystalline and crystalline structures. As temperature
increased to 120 and 173 ºC, these peaks began to broaden
and merge together. The broadening of spectral peaks is
commonly ascribed to melt behavior in polymeric materials.

Figure 4: Polypropylene recrystallization: G’ and G” (filled and open circles, respectively;

plotted on the left y-axis) and the normalized 808 cm-1 Raman shift peak height and the
3050-2800 cm-1 peak area (filled and open squares, respectively; plotted on the right
y-axis) as a function of decreasing temperature from 190 to 30 ºC.

Initially, at high temperatures (>150 ºC), G” was consistently

Figure 3: Raman spectra for polypropylene at 53 ºC (black), 120 ºC (yellow), and 173 ºC
(blue). The dashed line at 808 cm-1 and dashed box from 3050-2800 cm-1 indicate relevant greater than G’, indicating the PP specimen was in the melt
Raman bands to the recrystallization of polypropylene. state and displaying liquid-like behavior. As the temperature
decreased from 150 ºC to 100 ºC, an abrupt and drastic
Specific indicators in the Raman spectra for PP are the increase in both G’ and G” was observed at ~125 ºC. By
skeletal deformation of helical chains within the crystal the time the temperature reached 100 ºC, G’ had increased
(808 cm-1 peak) and the CH stretching region (3050-2800 by four orders of magnitude, while G” increased by two
cm-1 spectral range). The intensity of the 808 cm-1 peak will orders of magnitude. The temperature range from 150 to
be used as a measure of the crystallinity of PP2, and the 100 ºC can be viewed as the melt-to-crystalline transition
intensity in the CH stretching region is used as a measure region for this polypropylene material. As the temperature
of the overall Raman scattering intensity during the further decreased to below 100 ºC, G’ was significantly
crystallization process. For further analysis, the 808 cm-1 greater than G” by more than a full order of magnitude,
peak height was normalized by the peak area between 880 indicating the polypropylene had transitioned into the
and 780 cm-1. The 880-780 cm-1 spectral region contains semicrystalline and/or crystalline state, and was exhibiting
skeletal chain vibrations of all conformations during the solid-like behavior.
The observed plateau regions and increase in G’ and G” are Similar to the previous experiments, the rheology and
in direct agreement with the Raman spectral data (Figure Raman data were in agreement during the isothermal
4). In the melt region (~190-130 ºC)-1, the normalized 808 crystallization studies (Figures 5 and 6). The observed
cm-1 peak height was near zero and remained unchanged increase in the elastic and viscous moduli directly
by the decreasing temperature. However, as the PP melt correlates with the increase in the normalized 808 cm-1
began to crystallize, the normalized 808 cm-1 peak height peak height. Although not shown here, a minimum in the
significantly increased and was in unison with the abrupt CH stretching band intensity was again observed as the
increases in both G’ and G”. Also, as the PP sample started specimen transitioned from the liquid to the solid state.
to crystallize, the total peak area of the CH stretching region
decreased rapidly before any measurable growth of the 808
cm-1 band was observed. The decrease in intensity in the
3050-2800 cm-1 spectral region is postulated to be caused
by the formation of crystallites in the PP matrix. Crystallites
with dimensions on the same order of magnitude as the
wavelength of the incident laser (532 nm) would scatter
light, thereby creating a loss in the Raman signal intensity of
the CH stretching bands. The minimum in the CH stretching
profile was directly aligned with the observed crossover
between G’ and G” and the maximum slope of the
normalized 808 cm-1 peak profile. The correlation between
changes in the Raman spectral features and the rheological
response indicate that the crystallization rate was at its
maximum when the crystallite concentration was greatest.
Figure 5: Polypropylene isothermal recrystallization at 138 ºC : G’ and G” (filled and open
After the observed decrease in intensity, the integrated area
circles, respectively; plotted on the left y-axis) and the normalized 808 cm-1 Raman shift
of the CH stretching region returned to a value close to peak height (filled squares; plotted on the right y-axis) as a function of time.
what was detected in the melt phase.

In order to further evaluate the crystallization process of PP,

isothermal crystallization studies were performed at 150
and 138 ºC. The PP samples were heated at a minimum of
5 min at 190 ºC (to fully melt the specimen), rapidly cooled
at 10 ºC /min to 10 ºC above the crystallization temperature,
and then cooled at a slower rate of 2 ºC /min until the set
crystallization temperature was reached. The crystallization
temperature was then held constant for a maximum of 1
h (3600 s) to observe the isothermal crystallization of PP
as a function of time. Simultaneous rheology and Raman
spectroscopy data were acquired in-situ during crystallization
at 138 and 150 ºC (Figures 5 and 6, respectively). At
both isothermal temperatures, the PP material was initially
viscously dominated G’ < G” displaying liquid-like behavior. Figure 6: Polypropylene isothermal recrystallization at 150 ºC : G’ and G” (filled and open
circles, respectively; plotted on the left y-axis) and the normalized 808 cm-1 Raman shift
As time progressed, a crossover in G’ and G” was observed;
peak height (filled squares; plotted on the right y-axis) as a function of time.
where the moduli increased 2 to 3 orders of magnitude. The
crossover at 138 ºC occurred at 84 s into the isothermal
curing process, while it took 2150 s for the crossover to be
observed at 150 ºC. After the observed crossover, G’ and
G” reached a plateau and the crystallization process was
considered stabilized and complete.
Conclusions
A MARSXR system was employed to simultaneously measure the rheology and Raman spectroscopy of polypropylene
during the melt-to-crystallization transition process. This multimodal analytical tool allowed the bulk structural properties
of polypropylene (G’ and G”) to be directly correlated with conformational changes at the molecular level (helical chain
vibrations and CH stretching) in real-time. The data revealed that the rate of crystallization reached its maximum (indicated
by rapid increases in G’, G”, and the normalized 808 cm-1 peak height) when the crystallite concentration was greatest
(signified by a decrease in the overall Raman spectral intensity in the 3050-2800 cm-1 region). The observed correlation
between the macroscopic and molecular level measurements exemplifies the unique analytical capability unleashed by
hyphenating rheology with Raman spectroscopy. While this work focusses on the melting and crystallization of PP, the
underlying principles demonstrated here should be applicable for a wide range of material processes including gelation,
polymerization, curing behavior, as well as other shear-induced phenomena.

References
1. A.P. Kotula, M.W. Meyer, F. De Vito, J. Plog, A.R. Hight Walker and K.B. Migler, Review of Scientific Instruments, 87,
105105 (2016); DOI 10.1063/1.4903746.
2. R.A. Khafagy, Journal of Polymer Science: Part B: Polymer Physics, DOI 10.1002/polb

Find out more at thermofisher.com/mars

For Research Use Only. Not for use in diagnostic procedures. © 2017 Thermo Fisher Scientific Inc. All rights
reserved. All trademarks are the property of Thermo Fisher Scientific and its subsidiaries unless otherwise specified.
AN52959_E 06/17M
 APPLICATION NOTE AN53002

Investigating cocoa butter crystallization

using simultaneous rheology and Raman
spectroscopy (RheoRaman)
Author Application benefits
Nathan C. Crawford and Simultaneous rheology and Raman spectroscopy measurements were used to
Mohammed Ibrahim examine the isothermal crystallization of cocoa butter (CB). The results indicate
Thermo Fisher Scientific that CB crystallized by first hardening into an amorphous solid. The amorphous
Madison WI, USA solid then underwent a morphological transition to form a crystalline solid.
Without coupling these two separate analytical techniques, the observed
Keywords amorphous-solid to crystalline-solid transformation would have been left
Cocoa butter, crystallization, undetected. Alone, each technique suggests a single-stage process, however,
Raman spectroscopy, rheology, only when the two techniques are coupled is the multi-phase crystallization
RheoRaman, in situ, storage process revealed, further exemplifying the unique analytical capability
modulus G’, loss modulus G” unleashed by hyphenating rheology with in situ Raman spectroscopy.

Thermo Fisher Scientific Introduction

solutions Cocoa butter (CB) is an edible vegetable fat extracted from the cocoa
• HAAKE MARS 60 rheometer bean. CB is commonly used in home and personal care products (such as
ointments and lotions) and CB is a vital ingredient in chocolate. CB forms
• iXR Raman Spectrometer
the continuous phase within chocolate confections and is responsible for the
• HAAKE RheoRaman module chocolate’s texture, snap, gloss, melting behavior, and resistance to fat bloom.
These physical characteristics are a direct result of CB’s triacylglycerol (TAG)
• OMNIC for Dispersive Raman
composition and overall crystalline structure.
software

In general, TAG molecules assume a tuning fork configuration and the TAG
“forks” assemble to form crystal lattice structures. During crystallization,
the TAG molecules slow down as the CB oil cools and the TAGs come to
rest in contact with one another, forming what are known as “sub crystalline
cells.”1 Once the sub-cells are formed, they are thermodynamically driven to
aggregate into larger and more stable crystalline structures.2 The self-assembly
of sub-cell structures and their further aggregation is governed by a balance
of intra- and inter-molecular interactions. Depending on the molecular level
packing and orientation of the TAGs, CB can form different types of crystal
lattice structures (or polymorphs), where some crystal structures are more
desirable than others. Overall, CB crystallization is a highly complex, multistage
process. Understanding the isothermal crystallization behavior of CB is vital for
improving chocolate manufacturing processes and maintaining product quality.
In this note, rheology coupled with in situ Raman Rheology
spectroscopy was used to examine the isothermal Rheological measurements were performed using a
crystallization of cocoa butter. Raman spectroscopy Thermo Scientific™ HAAKE™ MARS 60 Rheometer
is a highly sensitive, relatively fast, and nondestructive equipped with a serrated 35 mm diameter plate rotor
technique that can probe the molecular structure and at a gap height of 1 mm. The serrated plate was
conformation in both liquid and solid TAG assemblies, used to prevent slip at the sample-rotor interface. All
as well as intra- and inter-TAG interactions. With measurements were conducted in the oscillatory mode,
simultaneous Raman spectra and rheological data, with a frequency of 1 Hz and a constant strain of 0.1%.
molecular-level interactions and conformational shifts CB samples were loaded onto the rheometer at 60 °C
during the isothermal crystallization of CB were directly and allowed to equilibrate for 10 min to erase any crystal
correlated with the changes in bulk viscoelastic structures and/or shear history from sample loading.
properties, providing unique insight into the multifaceted After the equilibrium step, the temperature was rapidly
crystallization behavior of cocoa butter. decreased from 60 °C to 22 °C at a rate of 10 °C/min.
The temperature was then held constant at 22 °C for
Materials and methods 120 min, collecting data every 10 s.
Materials
Commercially available, organic cocoa butter
(Theobroma cacao) was acquired from Inesscents™
(Ashland, OR, USA).

A B

Figure 1. (a) The Thermo Scientific™ HAAKE™ MARSXR RheoRaman System. (b)
Schematic diagram of the MARSXR RheoRaman system (showing side and top views of
the rheometer sample stage). The iXR Raman spectrometer is free-space coupled to the
MARS rheometer using lens tubes and mirrors that direct light into a 20x objective. The
objective focuses the incoming laser (green dashed line) and collects the back-scattered
Raman light (yellow) coming out of the sample sitting atop the rheometer stage.

Raman spectroscopy RheoRaman coupling

Raman spectroscopy measurements were performed The Thermo Scientific™ HAAKE™ MARSXR RheoRaman
using a Thermo Scientific™ iXR™ Raman Spectrometer. System consists of the iXR Raman spectrometer and the
A 532 nm laser was used with 10 mW laser power at HAAKE MARS 60 rheometer coupled together using the
the sample. The spectral range was 50-3500 cm-1. The HAAKE RheoRaman module (Figure 1a). The iXR Raman
spectra were collected using a 2-second exposure spectrometer was free-space coupled to the rheometer
time and 4 sample exposures. Data acquisition and with an optical train which used a series of mirrors to direct
processing were controlled by the Thermo Scientific™ the incident laser into the RheoRaman module (Figure 1b).
OMNIC™ Software for Dispersive Raman. For the data Within the module, a mirror directed the laser beam into
presented here, Sequential Raman spectra (in parallel a 20x objective, where the laser light was focused into
with the Rheological measurements) were collected the sample (perpendicularly to the flow or vorticity plane).
over a predetermined time window using the time series Backscattered Raman light was collected using the same
collection function of the SERIES software within the 20x objective and guided back to the spectrometer using
OMNIC for Dispersive Raman software package. the same optical train as the incident laser (eventually to
the spectrograph inside the spectrometer; Figure 1b).
The sample was positioned between a sandblasted glass stretching region (1700-1800 cm-1), the olefinic (C=C)
bottom plate and the serrated 35 mm plate rotor (the band at ~1655 cm-1, the CH3 and CH2 deformations
textured plates were used to avoid slip at the sample- (~1460 and 1440 cm-1, respectively), the CH2 twisting
plate interfaces). An electrical heating element within region (1250-1300 cm-1), and the C–C stretching region
the RheoRaman module provided temperature control (1000-1200 cm-1).
from below the sample, while an active electrical hood
was used to provide temperature control from above The C-H stretching regions for the melted and solidified
(eliminating the potential for a temperature gradient within CB specimens are highlighted in Figure 3. Two strong
the sample). Cooling of the sample was supplied from a peaks were observed at ~2850 cm-1 and 2882 cm-1,
temperature-controlled water bath circulator. which are attributed to symmetric and asymmetric CH2
stretching, respectively.2 The symmetric vibrational modes
Results and discussion at 2850 cm-1 were dominant in the liquid (melt) phase, while
Raman spectroscopy: Cocoa butter crystallization the asymmetric vibrations at 2882 cm-1 were dominant in
Raman spectra for the liquid phase CB melt and the the solid phase. Thus, the 2850 cm-1 and 2882 cm-1 bands
crystalline solid CB in the 500-3100 cm-1 range are shown are strong indicators of amorphous and crystalline content,
in Figure 2. Prominent Raman features were observed in respectively.3 Subsequently, the I2882/I2850 peak intensity
both the C–H stretching region (2700-3050 cm-1) and the ratio was used to dynamically track crystal formation
fingerprint region (1000-1800 cm-1). More specifically, the during the in situ RheoRaman measurements.
lower Raman shift features include: the carbonyl (C=O)

Figure 2. The full Raman spectra of melted

and crystalline cocoa butter.

Figure 3. Raman spectra of the C-H stretching

region (2700-3050 cm-1) for melted and
crystalline cocoa butter.

Although less intense than the C-H stretching region, Two well-defined features emerged at 1130 cm-1 and
approximately eight unique spectral features were 1063 cm-1 during the solidification process, which
identified in the fingerprint region (1000-1800 cm-1; originate from the symmetric and asymmetric C-C
Figure 4). When comparing the CB melt state to the stretching, respectively.4,5 In the melt phase, all C-C
crystalline phase, the most significant changes were stretching bands were relatively weak and broad due to
observed in the C–C stretching region (1000-1200 cm-1). the disordering effects of methyl gauche conformations.
 Figure 4. The 1000–1800 cm-1 Raman spectral
range for melted and crystalline cocoa butter.

However, as the CB solidified, the backbone methyl Unlike the individual moduli, tan(δ) can be used to quantify
groups were ordered into the trans-conformation, overall brittleness of a material and is commonly used
signified by the emergence of the peak at 1130 cm-1. to assess glass transition behavior. In general, as tan(δ)
Therefore, in addition to the I2882/I2850 peak intensity ratio, becomes smaller, the more G’ deviates from G”, and the
the I1130/I2850 spectral marker was also used to track more brittle (or glass-like) the material becomes.
the crystalline-phase transition within CB via in situ
rheoRaman measurements. During the initial portion of the isothermal hold at 22 °C
from 0 to 5 min (immediately following the rapid decrease
Simultaneous rheology and Raman spectroscopy in temperature from 60 °C to 22 °C), both G’ and G”
(RheoRaman) increased as the CB transformed from a melted liquid to a
The melt-to-solid phase transition of cocoa butter was soft semi-solid (Figure 5a). This initial increase in modulus
probed rheologically using small amplitude oscillatory is most likely due to a delay between the set temperature
shear measurements (Figure 5a), where the storage and the internal temperature of the loaded sample. Once
modulus G’ and loss modulus G” were measured as a the sample had reached thermal equilibrium and was
function of time at the isothermal temperature of 22 °C. at the isothermal set point of 22 °C, the moduli were
G’ and G” are measures of a material’s elastic and relatively stable from 10 to 25 min. From 25 to 50 min,
viscous behavior, respectively. A liquid-like material will however, both G’ and G” begin to gradually increase and
be more viscous than elastic (i.e., viscously dominated), then from 50 to 80 min, the moduli rapidly increased,
and as a result, G” will be greater than G’. Conversely, a where G’ and G” increased by approximately 5 and
solid-like material will display more elastic than viscous 4 orders of magnitude, respectively. The exponential
behavior (i.e., elastically dominated), where G’ will be increase in the moduli indicates a solidification process,
greater than G”. The overall magnitudes of G’ and G”, where the CB transformed from a pliable semi-solid to
as well as their relative difference in magnitude, often a more robust, hardened solid. At 80 min and beyond,
reported as the ratio of G”/G’, determines the general growth in the elastic modulus slowed and eventually
viscoelasticity and overall resistance to deformation for a plateaued, showing no further significant change past
given material. 100 min. The viscous modulus, however, reached a
slight plateau from 80 to 100 min and then proceeded to
The ratio of G”/G’ (plotted on the right y-axis of Figure 5a) gradually decrease from 100 min and beyond.
is commonly used to track viscoelasticity of a material:
During the increase in G’ and G”, a rapid decrease in
the loss factor tan(δ) was observed from ~65 min and
beyond (Figure 5a, right y-axis). The decrease in the loss
where δ is the phase angle defined as the shift or lag factor indicates a deviation in overall magnitude between
between the input strain and resultant stress sine waves G’ and G”. As the CB hardened, the increase in G’
(or vice versa) during an oscillatory shear measurement. exceeded the increase in G”, triggering the decrease in
The term “tan(δ)” is often referred to as the loss or damping tan(δ). At the end of the 120 min isothermal study, G’ was
factor. Values of tan(δ) less than unity indicate elastically more than a full order of magnitude greater than G” and
dominant (solid-like) behavior, while values greater than the loss factor was approaching 0.01, indicating the CB
unity indicate viscously dominant (liquid-like) behavior. had transitioned into a brittle glass-like solid.
 Raman spectral markers, on the other hand, are indicators
A of crystal formation. Thus, the time delay between the
rheology and Raman profiles suggests that CB first hardens
into an amorphous solid, followed by a transformation from
an amorphous to a crystalline solid. This morphological
transformation was signified by the subsequent increase
in the Raman band intensities associated with crystal CB
structures (the 1130 and 2882 cm-1 peaks). The temporal
separation of the rheological and Raman spectral profiles
indicates a clear distinction between bulk hardening of the
CB and the formation of crystalline domains.

Interestingly, the increase in the Raman spectral features

(I1130/I2850 and I2882/I2850) directly correlated with the observed
B reduction in tan(δ) (Figure 5a and b). The loss factor is an
indication of material brittleness and crystalline structures
are commonly known to be brittle. Thus, it is reasonable that
the formation of crystal domains at the molecular level (as
indicated by Raman) coincides with the overall brittleness of
the CB. As a result, the loss factor may be a more revealing
indicator of bulk CB crystallization than G’ and G” alone.

Conclusions
Simultaneous rheology and Raman spectroscopy
measurements were used to examine the isothermal
crystallization of cocoa butter. This multimodal analytical
technique allowed the bulk mechanical properties of cocoa
Figure 5. (a) Rheology: G’ and G” (filled and open circles, respectively;
plotted on the left y-axis) and tan(δ) (plotted on the right y-axis) and butter (G’, G”, and tan(δ)) to be directly correlated with
(b) Raman: the I1130 /I2850 (left y-axis, green) and I2882/I2850 (right y-axis, conformational changes at the molecular level (νas(CH2)
black) peak intensity ratios for CB during isothermal crystallization
mode at 2882 cm-1 and the νs(C-C) mode at 1130 cm-1)
at 22 °C. The vertical dashed line at 45 min indicates the increase of
G’ and G”, while the dashed line at 65 min indicates the decrease in in real-time. After rapid cooling (10 °C /min) and at an
tan(δ) and increase in the Raman ratios. isothermal temperature of 22 °C, there was a noticeable
time lag between the rheological response (G’ and G”)
The observed rheological behavior was further confirmed and the Raman spectral profiles. The observed time delay
using simultaneous Raman spectroscopy (Figure 5b). indicates that CB crystallized by first hardening into an
Initially, both the I1130/I2850 and I2882/I2850 peak intensity amorphous solid, manifested by a sharp increase in G’ and
ratios remained unchanged during the first ~65 min of the G” while the Raman features remained unchanged. The
isothermal study. Then a sharp increase of the I1130/I2850 and amorphous solid then underwent a morphological transition
I2882/I2850 ratios began at ~65 min, indicating the formation to form a crystalline solid, signified by the increase in
of crystal structures within the CB. As the CB further Raman features associated with crystal CB structures
crystallized, both spectral markers continued to increase (1130 and 2882 cm-1). Without coupling these two separate
from 65 to 100 min. Beyond 100 min, the growth in both analytical techniques, the observed amorphous-solid
Raman features had subsided and the peak intensity ratios to crystalline-solid transformation would have been left
began to stabilize. undetected. Alone, each technique suggests a single-
stage process, however, only when the two techniques are
Overall, the rate of increase in the 1130 and 2882 cm-1 coupled is the multi-phase crystallization process revealed,
spectral ratios were similar to the rate of change for both G’ further exemplifying the unique analytical capability
and G” (i.e., they increased with similar slopes). However, unleashed by hyphenating rheology with in situ Raman
there was a noticeable 15-20 min lag between the observed spectroscopy. While this work focusses on the isothermal
increase in G’ and G” and the rise of the Raman intensity crystallization of CB, the underlying principles applied here
ratios. The sharp upturn in G’ and G” indicates an increased should be applicable for a wide range of material processes
resistance to deformation (i.e., a bulk hardening of the including gelation, polymerization, curing behavior, as well
CB), signaling the start of the solidification process. The as other shear-induced phenomena.
References
1. K. Sato, Crystallization of Lipids: Fundamentals and
Applications in Food, Cosmetics, and Pharmaceuti-
cals. Hoboken, NJ: John Wiley & Sons, 2018.
2. S. Bresson, D. Rousseau, S. Ghosh, M. El Marssi,
and V. Faivre, Raman spectroscopy of the polymor-
phic forms and liquid state of cocoa butter, Eur. J.
Lipid Sci. Technol. 113, 992–1004, 2011.
3. R. G. Snyder, H. L. Strauss, and C. A. Elliger, C–H
stretching modes and the structure of n-alkyl chains.
1. Long, disordered chains, J. Phys. Chem. 86,
5145–5150, 1982.
4. R. J. Meier, Studying the length of trans confor-
mational sequences in polyethylene using Raman
spectroscopy: A computational study, Polymer. 43,
517–522, 2002.
5. M. Zheng and M. Du, Phase behavior, conformations,
thermodynamic properties, and molecular motion of
multicomponent paraffin waxes: A Raman spectros-
copy study. Vib. Spectrosc. 40, 219–224, 2006.

Find out more at www.thermofisher.com/Raman

©2018 Thermo Fisher Scientific Inc. All trademarks are the property of Thermo Fisher Scientific and its
subsidiaries unless otherwise specified. AN53002_E 03/18M
Other resources
Basic aspects of experimental design in
Raman spectroscopy
Author
Alexander Rzhevskii • Thermo Fisher Scientific, Madison WI, USA

Abstract
This article discusses basic practical aspects of analytical instrument performance and
experimental design that should be taken into consideration when characterizing microscale
objects using a Raman microscope. Proper instrument alignment, optical objective magnification,
confocal aperture, and sampling step settings must ensure sufficient spatial resolution and
measurement precision to discriminate a Raman signal of the object from a surrounding matrix.
The essential relationships between the spectral measurement parameters are considered
theoretically and illustrated experimentally.

Read online ›
Spectroscopy, 2016, Volume 31, Issue 11, p 40–45

Resonance Raman probes for organelle-specific

labeling in live cells
Author
Andrey N. Kuzmin1, Artem Pliss1, Chang-Keun Lim1,2, Jeongyun Heo1, Sehoon Kim1,2, Alexander
Rzhevskii3, Bobo Gu1,4, Ken-Tye Yong5, Shangchun Wen4 & Paras N. Prasad1
1 Institute for Lasers, Photonics and Biophotonics, State University of New York, University at Buffalo, Buffalo, NY, 14260, USA.
2 Center for Theragnosis, Korea Institute of Science and Technology, Seoul 136-791, Republic of Korea.
3 Thermo Fisher Scientific, Waltham, MA 02451, USA.
4 Key Laboratory for Micro-/Nano-Optoelectronic Devices of Ministry of Education, School of Physics and Electronics, Hunan University,
Changsha 410082, China.

5 School of Electrical and Electronic Engineering, Nanyang Technological University, 639798, Singapore.

Abstract
Raman microspectroscopy provides for high-resolution non-invasive molecular analysis
of biological samples and has a breakthrough potential for dissection of cellular molecular
composition at a single organelle level. However, the potential of Raman microspectroscopy
can be fully realized only when novel types of molecular probes distinguishable in the Raman
spectroscopy modality are developed for labeling of specific cellular domains to guide
spectrochemical spatial imaging. Here we report on the design of a next generation Raman
probe, based on BlackBerry Quencher 650 compound, which provides unprecedentedly high
signal intensity through the Resonance Raman (RR) enhancement mechanism. Remarkably,
RR enhancement occurs with low-toxic red light, which is close to maximum transparency in
the biological optical window. The utility of proposed RR probes was validated for targeting
lysosomes in live cultured cells, which enabled identification and subsequent monitoring of
dynamic changes in this organelle by Raman imaging.

Read online ›
Sci. Rep. 6, 28483; doi: 10.1038/srep28483 (2016)
 Tracking microplastics in the environment via
FTIR microscopy
Authors
Michael Bradley, Suja Sukumaran, Steven Lowry, Stephan Woods
Thermo Fisher Scientific, Madison WI, USA

Abstract
Microplastics are particulates, roughly 20–1000 μm in size, originating from materials such as
clothing, abrasive action on plastics, or engineered microbeads as found in some exfoliating
cosmetics. The microplastics enter aquifers where the particles can be consumed by filter
feeders. Microplastics are chemically stable, giving them a long lifetime in the environment and
making excretion or digestion difficult. Analytically, the size and polymeric nature of microplastics
makes Fourier transform infrared (FTIR) microscopy an ideal tool for detection and identification.
Standard analyses typically start with a filtration step, extracting the material from the matrix. The
analysis can proceed directly on the dried filter without further sample preparation. This simplicity
in both sampling and analysis enables the rapid assessment of microplastic encroachment and
can assist in the development of remediation techniques. We show examples from both prepared
and field samples using microattenuated total reflection (ATR) FTIR.

Read online ›
Spectroscopy, 2017, Special Issues Volume 32, Issue 8, p 17–23

Find out more at www.thermofisher.com/ftir

For Research Use Only. Not for use in diagnostic procedures. ©2018 Thermo Fisher Scientific Inc.
All trademarks are the property of Thermo Fisher Scientific and its subsidiaries unless otherwise specified.
BR53012_E 05/18M